{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92cf1665bb8efa753f36","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ff43ab3710001fb63ba2cbba","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.07978474570120461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1158,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[578];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1158,"metadata":{"avgScore":0.07978474570120461,"avg_score":0.07978474570120461,"avg_score_percent":7.978474570120461,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":14.4,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":84561224,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":266098,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":48999,"source_flags":[],"source_harness":"codex","source_row_index":578,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":14.4,"totalDurationS":48999,"totalInputTokens":84561224,"totalOutputTokens":266098,"total_cost_usd":14.4,"total_input_tokens":84561224,"total_output_tokens":266098,"total_runtime_s":48999,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.978474570120461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4a9f6df6db1bbb66b086","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ff2f5ab1861f9ab70c83e8e0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5206445104262789,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1328,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[663];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1328,"metadata":{"avgScore":0.5206445104262789,"avg_score":0.5206445104262789,"avg_score_percent":52.064451042627894,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":212.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1305756331,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":3877107,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":46,"runs":141,"runtime":189252,"source_flags":[],"source_harness":"codex","source_row_index":663,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":212.65,"totalDurationS":189252,"totalInputTokens":1305756331,"totalOutputTokens":3877107,"total_cost_usd":212.65,"total_input_tokens":1305756331,"total_output_tokens":3877107,"total_runtime_s":189252,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.064451042627894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a9e5f56d5e0918bfeabc","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f1fdeeb18f493765ec4a03b6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.15451296049520938,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":648,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[323];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":648,"metadata":{"avgScore":0.15451296049520938,"avg_score":0.15451296049520938,"avg_score_percent":15.451296049520938,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.47,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":129505530,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":285267,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":48790,"source_flags":[],"source_harness":"codex","source_row_index":323,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":22.47,"totalDurationS":48790,"totalInputTokens":129505530,"totalOutputTokens":285267,"total_cost_usd":22.47,"total_input_tokens":129505530,"total_output_tokens":285267,"total_runtime_s":48790,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":15.451296049520938,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0b4b05800cbb3308acc1","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f185b5bd79c17d6d982d58b5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6736742823508564,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":388,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[193];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":388,"metadata":{"avgScore":0.6736742823508564,"avg_score":0.6736742823508564,"avg_score_percent":67.36742823508564,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":228811254,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1065059,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":67,"runtime":60584,"source_flags":[],"source_harness":"codex","source_row_index":193,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":46.18,"totalDurationS":60584,"totalInputTokens":228811254,"totalOutputTokens":1065059,"total_cost_usd":46.18,"total_input_tokens":228811254,"total_output_tokens":1065059,"total_runtime_s":60584,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.36742823508564,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5fcb4b781f286a1bb8c0","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e5b9e54d80a53ac4621a920c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5206445104262789,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":856,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[427];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":856,"metadata":{"avgScore":0.5206445104262789,"avg_score":0.5206445104262789,"avg_score_percent":52.064451042627894,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":212.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1305756331,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":3877107,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":46,"runs":141,"runtime":189252,"source_flags":[],"source_harness":"codex","source_row_index":427,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":212.65,"totalDurationS":189252,"totalInputTokens":1305756331,"totalOutputTokens":3877107,"total_cost_usd":212.65,"total_input_tokens":1305756331,"total_output_tokens":3877107,"total_runtime_s":189252,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.064451042627894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d47a5969213a672ed34","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e002fa0b7775cda6b162ba3b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4372134877427408,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":136,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[67];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":136,"metadata":{"avgScore":0.4372134877427408,"avg_score":0.4372134877427408,"avg_score_percent":43.721348774274084,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.31,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":202023719,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":828678,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":15,"runs":55,"runtime":37502,"source_flags":[],"source_harness":"codex","source_row_index":67,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":38.31,"totalDurationS":37502,"totalInputTokens":202023719,"totalOutputTokens":828678,"total_cost_usd":38.31,"total_input_tokens":202023719,"total_output_tokens":828678,"total_runtime_s":37502,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.721348774274084,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c70142e0bf26c4ceab78","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-de19737888e5d736334e5a13","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.507344624718011,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":726,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[362];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":726,"metadata":{"avgScore":0.507344624718011,"avg_score":0.507344624718011,"avg_score_percent":50.7344624718011,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":143.63,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":910676386,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2836040,"passRate":0.29523809523809524,"pass_rate":0.29523809523809524,"pass_rate_percent":29.523809523809526,"passes":31,"runs":105,"runtime":89570,"source_flags":[],"source_harness":"codex","source_row_index":362,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":143.63,"totalDurationS":89570,"totalInputTokens":910676386,"totalOutputTokens":2836040,"total_cost_usd":143.63,"total_input_tokens":910676386,"total_output_tokens":2836040,"total_runtime_s":89570,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.7344624718011,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6477dd68a5fe12894f6b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-dddb41967a7b16ff8d81ef36","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5331379784041669,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":848,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[423];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":848,"metadata":{"avgScore":0.5331379784041669,"avg_score":0.5331379784041669,"avg_score_percent":53.31379784041669,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":371.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2398634685,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":6446026,"passRate":0.3120567375886525,"pass_rate":0.3120567375886525,"pass_rate_percent":31.20567375886525,"passes":44,"runs":141,"runtime":199793,"source_flags":[],"source_harness":"codex","source_row_index":423,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":371.53,"totalDurationS":199793,"totalInputTokens":2398634685,"totalOutputTokens":6446026,"total_cost_usd":371.53,"total_input_tokens":2398634685,"total_output_tokens":6446026,"total_runtime_s":199793,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.31379784041669,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-737c4b194a049e67b1bc","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d9b46f9679c8d135f42b4a79","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40056329147969094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1384,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[691];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1384,"metadata":{"avgScore":0.40056329147969094,"avg_score":0.40056329147969094,"avg_score_percent":40.056329147969095,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.72,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":223199374,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":996976,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":140221,"source_flags":[],"source_harness":"codex","source_row_index":691,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":44.72,"totalDurationS":140221,"totalInputTokens":223199374,"totalOutputTokens":996976,"total_cost_usd":44.72,"total_input_tokens":223199374,"total_output_tokens":996976,"total_runtime_s":140221,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.056329147969095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cd5f33531723b06360fe","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d438de9ba7bcee6b5368d978","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.49414814826021464,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":494,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[246];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":494,"metadata":{"avgScore":0.49414814826021464,"avg_score":0.49414814826021464,"avg_score_percent":49.41481482602146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235.12,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1435261861,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":4162374,"passRate":0.3026315789473684,"pass_rate":0.3026315789473684,"pass_rate_percent":30.263157894736842,"passes":46,"runs":152,"runtime":238042,"source_flags":[],"source_harness":"codex","source_row_index":246,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":235.12,"totalDurationS":238042,"totalInputTokens":1435261861,"totalOutputTokens":4162374,"total_cost_usd":235.12,"total_input_tokens":1435261861,"total_output_tokens":4162374,"total_runtime_s":238042,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.41481482602146,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3ac5d31994eec7ef3729","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d2de271165f71939c3e85403","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7562644213793285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":362,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[180];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":362,"metadata":{"avgScore":0.7562644213793285,"avg_score":0.7562644213793285,"avg_score_percent":75.62644213793284,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.71,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":359705378,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1633460,"passRate":0.4925373134328358,"pass_rate":0.4925373134328358,"pass_rate_percent":49.25373134328358,"passes":33,"runs":67,"runtime":69711,"source_flags":[],"source_harness":"codex","source_row_index":180,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":68.71,"totalDurationS":69711,"totalInputTokens":359705378,"totalOutputTokens":1633460,"total_cost_usd":68.71,"total_input_tokens":359705378,"total_output_tokens":1633460,"total_runtime_s":69711,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":75.62644213793284,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c1903d0d3f40aadab81b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ce887a04b39625cb59a09ee9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.11752938705842898,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1120,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[559];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1120,"metadata":{"avgScore":0.11752938705842898,"avg_score":0.11752938705842898,"avg_score_percent":11.7529387058429,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":94.89,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":665277010,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1151373,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":71143,"source_flags":[],"source_harness":"codex","source_row_index":559,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":94.89,"totalDurationS":71143,"totalInputTokens":665277010,"totalOutputTokens":1151373,"total_cost_usd":94.89,"total_input_tokens":665277010,"total_output_tokens":1151373,"total_runtime_s":71143,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.7529387058429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c74400bcabaa4905eda","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ca34cd5bfb5f1e0052606a68","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.49414814826021464,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":22,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[10];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":22,"metadata":{"avgScore":0.49414814826021464,"avg_score":0.49414814826021464,"avg_score_percent":49.41481482602146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235.12,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1435261861,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":4162374,"passRate":0.3026315789473684,"pass_rate":0.3026315789473684,"pass_rate_percent":30.263157894736842,"passes":46,"runs":152,"runtime":238042,"source_flags":[],"source_harness":"codex","source_row_index":10,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":235.12,"totalDurationS":238042,"totalInputTokens":1435261861,"totalOutputTokens":4162374,"total_cost_usd":235.12,"total_input_tokens":1435261861,"total_output_tokens":4162374,"total_runtime_s":238042,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.41481482602146,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d32b177f28e74672a68f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c40c2b5ded82b4717e0f1024","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10993413212460641,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":288,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[143];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":288,"metadata":{"avgScore":0.10993413212460641,"avg_score":0.10993413212460641,"avg_score_percent":10.993413212460641,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.64,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":284905745,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":772005,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":75483,"source_flags":[],"source_harness":"codex","source_row_index":143,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":46.64,"totalDurationS":75483,"totalInputTokens":284905745,"totalOutputTokens":772005,"total_cost_usd":46.64,"total_input_tokens":284905745,"total_output_tokens":772005,"total_runtime_s":75483,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.993413212460641,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f0d2b3091559949ea8f1","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c38ce812ed78e9c3c87fe010","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40056329147969094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":912,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[455];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":912,"metadata":{"avgScore":0.40056329147969094,"avg_score":0.40056329147969094,"avg_score_percent":40.056329147969095,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.72,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":223199374,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":996976,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":140221,"source_flags":[],"source_harness":"codex","source_row_index":455,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":44.72,"totalDurationS":140221,"totalInputTokens":223199374,"totalOutputTokens":996976,"total_cost_usd":44.72,"total_input_tokens":223199374,"total_output_tokens":996976,"total_runtime_s":140221,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.056329147969095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-adde5dd1a79b35177bde","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bd39df45b225832ffe65397e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4253164625672212,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":144,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[71];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":144,"metadata":{"avgScore":0.4253164625672212,"avg_score":0.4253164625672212,"avg_score_percent":42.53164625672212,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.34,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":652436878,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2206867,"passRate":0.2909090909090909,"pass_rate":0.2909090909090909,"pass_rate_percent":29.09090909090909,"passes":16,"runs":54,"runtime":71516,"source_flags":[],"source_harness":"codex","source_row_index":71,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":107.34,"totalDurationS":71516,"totalInputTokens":652436878,"totalOutputTokens":2206867,"total_cost_usd":107.34,"total_input_tokens":652436878,"total_output_tokens":2206867,"total_runtime_s":71516,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.53164625672212,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f003eb61bf800613c8e5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bd2929b576735cefd268d960","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5037352082410917,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":20,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[9];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":20,"metadata":{"avgScore":0.5037352082410917,"avg_score":0.5037352082410917,"avg_score_percent":50.37352082410918,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":390.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2499272937,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":6752715,"passRate":0.2894736842105263,"pass_rate":0.2894736842105263,"pass_rate_percent":28.947368421052634,"passes":44,"runs":149,"runtime":239192,"source_flags":[],"source_harness":"codex","source_row_index":9,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":390.61,"totalDurationS":239192,"totalInputTokens":2499272937,"totalOutputTokens":6752715,"total_cost_usd":390.61,"total_input_tokens":2499272937,"total_output_tokens":6752715,"total_runtime_s":239192,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.37352082410918,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-57caf409039ccf13036f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bbeda83374aef769ba0264ab","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.46128766882143674,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":34,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[16];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":34,"metadata":{"avgScore":0.46128766882143674,"avg_score":0.46128766882143674,"avg_score_percent":46.12876688214367,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.52,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":699354083,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2559522,"passRate":0.24342105263157895,"pass_rate":0.24342105263157895,"pass_rate_percent":24.342105263157894,"passes":37,"runs":152,"runtime":165192,"source_flags":[],"source_harness":"codex","source_row_index":16,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":127.52,"totalDurationS":165192,"totalInputTokens":699354083,"totalOutputTokens":2559522,"total_cost_usd":127.52,"total_input_tokens":699354083,"total_output_tokens":2559522,"total_runtime_s":165192,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.12876688214367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ae41cac005d1b3d5c7c8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b444f56c8cfaff3a5528bf15","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3361045019913022,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":804,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[401];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":804,"metadata":{"avgScore":0.3361045019913022,"avg_score":0.3361045019913022,"avg_score_percent":33.61045019913022,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.85,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":56323564,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":407087,"passRate":0.11428571428571428,"pass_rate":0.11428571428571428,"pass_rate_percent":11.428571428571429,"passes":12,"runs":105,"runtime":33091,"source_flags":[],"source_harness":"codex","source_row_index":401,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":12.85,"totalDurationS":33091,"totalInputTokens":56323564,"totalOutputTokens":407087,"total_cost_usd":12.85,"total_input_tokens":56323564,"total_output_tokens":407087,"total_runtime_s":33091,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":33.61045019913022,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a4b13f646c3115e7ec7c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ae8990cc48b17bb9bb8e32df","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.46128766882143674,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":506,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[252];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":506,"metadata":{"avgScore":0.46128766882143674,"avg_score":0.46128766882143674,"avg_score_percent":46.12876688214367,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.52,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":699354083,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2559522,"passRate":0.24342105263157895,"pass_rate":0.24342105263157895,"pass_rate_percent":24.342105263157894,"passes":37,"runs":152,"runtime":165192,"source_flags":[],"source_harness":"codex","source_row_index":252,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":127.52,"totalDurationS":165192,"totalInputTokens":699354083,"totalOutputTokens":2559522,"total_cost_usd":127.52,"total_input_tokens":699354083,"total_output_tokens":2559522,"total_runtime_s":165192,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.12876688214367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c922c73be65a5f53c9d4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ac68eeee2b918ed7dddb14a4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4177055625187485,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":982,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[490];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":982,"metadata":{"avgScore":0.4177055625187485,"avg_score":0.4177055625187485,"avg_score_percent":41.77055625187485,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":53.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":299155581,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1250311,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":14,"runs":50,"runtime":51596,"source_flags":[],"source_harness":"codex","source_row_index":490,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":53.18,"totalDurationS":51596,"totalInputTokens":299155581,"totalOutputTokens":1250311,"total_cost_usd":53.18,"total_input_tokens":299155581,"total_output_tokens":1250311,"total_runtime_s":51596,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.77055625187485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-55240582e48140c96575","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-abca6778d092f9e791cbe141","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2343520785550317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":610,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[304];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":610,"metadata":{"avgScore":0.2343520785550317,"avg_score":0.2343520785550317,"avg_score_percent":23.435207855503172,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":11.06,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":56109934,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":143436,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":22844,"source_flags":[],"source_harness":"codex","source_row_index":304,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":11.06,"totalDurationS":22844,"totalInputTokens":56109934,"totalOutputTokens":143436,"total_cost_usd":11.06,"total_input_tokens":56109934,"total_output_tokens":143436,"total_runtime_s":22844,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.435207855503172,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c4435861018c6c37e014","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-aa3cb750ae69a1a34c7f42f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3280611890351934,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1402,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[700];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1402,"metadata":{"avgScore":0.3280611890351934,"avg_score":0.3280611890351934,"avg_score_percent":32.80611890351934,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":102309910,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":562218,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":141,"runtime":77880,"source_flags":[],"source_harness":"codex","source_row_index":700,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":22.46,"totalDurationS":77880,"totalInputTokens":102309910,"totalOutputTokens":562218,"total_cost_usd":22.46,"total_input_tokens":102309910,"total_output_tokens":562218,"total_runtime_s":77880,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.80611890351934,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d67c74145ea8bd4f3845","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a810fd86eb7e27f4779ab497","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2963635563049718,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1036,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[517];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1036,"metadata":{"avgScore":0.2963635563049718,"avg_score":0.2963635563049718,"avg_score_percent":29.636355630497178,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":11.82,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":51495533,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":311061,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":50,"runtime":27012,"source_flags":[],"source_harness":"codex","source_row_index":517,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":11.82,"totalDurationS":27012,"totalInputTokens":51495533,"totalOutputTokens":311061,"total_cost_usd":11.82,"total_input_tokens":51495533,"total_output_tokens":311061,"total_runtime_s":27012,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.636355630497178,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b175296c3485af11933b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a5de15466786bf2f839f4c8e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.09622840053229612,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1138,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[568];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1138,"metadata":{"avgScore":0.09622840053229612,"avg_score":0.09622840053229612,"avg_score_percent":9.622840053229611,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":5.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":29419495,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":127636,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":14261,"source_flags":[],"source_harness":"codex","source_row_index":568,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":5.65,"totalDurationS":14261,"totalInputTokens":29419495,"totalOutputTokens":127636,"total_cost_usd":5.65,"total_input_tokens":29419495,"total_output_tokens":127636,"total_runtime_s":14261,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.622840053229611,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-caa30c6bfec175611dfe","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a387939fd75e48a98193c9fc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10970741091203197,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":290,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[144];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":290,"metadata":{"avgScore":0.10970741091203197,"avg_score":0.10970741091203197,"avg_score_percent":10.970741091203196,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":109.66,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":742917153,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1343720,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":108705,"source_flags":[],"source_harness":"codex","source_row_index":144,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":109.66,"totalDurationS":108705,"totalInputTokens":742917153,"totalOutputTokens":1343720,"total_cost_usd":109.66,"total_input_tokens":742917153,"total_output_tokens":1343720,"total_runtime_s":108705,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.970741091203196,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ea09a9cefa1da2f6cc41","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-99ca04c05184c4be617585b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3007389590993101,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":196,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[97];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":196,"metadata":{"avgScore":0.3007389590993101,"avg_score":0.3007389590993101,"avg_score_percent":30.073895909931007,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":18.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":85588559,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":355569,"passRate":0.12727272727272726,"pass_rate":0.12727272727272726,"pass_rate_percent":12.727272727272727,"passes":7,"runs":55,"runtime":32130,"source_flags":[],"source_harness":"codex","source_row_index":97,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":18.13,"totalDurationS":32130,"totalInputTokens":85588559,"totalOutputTokens":355569,"total_cost_usd":18.13,"total_input_tokens":85588559,"total_output_tokens":355569,"total_runtime_s":32130,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.073895909931007,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-25521a18ef4537464f03","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-96f18889d45a18886f363ed0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1505198327191624,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1096,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[547];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1096,"metadata":{"avgScore":0.1505198327191624,"avg_score":0.1505198327191624,"avg_score_percent":15.051983271916239,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":189.81,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1321242146,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2441114,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":30,"runtime":79440,"source_flags":[],"source_harness":"codex","source_row_index":547,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":189.81,"totalDurationS":79440,"totalInputTokens":1321242146,"totalOutputTokens":2441114,"total_cost_usd":189.81,"total_input_tokens":1321242146,"total_output_tokens":2441114,"total_runtime_s":79440,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":15.051983271916239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-79e9f321f28666c717a0","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-961a568bf379e956fd97f8a3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4789918638067592,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":870,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[434];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":870,"metadata":{"avgScore":0.4789918638067592,"avg_score":0.4789918638067592,"avg_score_percent":47.899186380675914,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":116.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":643244149,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2416086,"passRate":0.2624113475177305,"pass_rate":0.2624113475177305,"pass_rate_percent":26.24113475177305,"passes":37,"runs":141,"runtime":142348,"source_flags":[],"source_harness":"codex","source_row_index":434,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":116.46,"totalDurationS":142348,"totalInputTokens":643244149,"totalOutputTokens":2416086,"total_cost_usd":116.46,"total_input_tokens":643244149,"total_output_tokens":2416086,"total_runtime_s":142348,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.899186380675914,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-15078c14d412c64db075","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-91ab6bf0373a9d75a8791d22","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.48273103955698327,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":446,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[222];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":446,"metadata":{"avgScore":0.48273103955698327,"avg_score":0.48273103955698327,"avg_score_percent":48.273103955698325,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":10.56,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":46726983,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":263717,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":67,"runtime":50813,"source_flags":[],"source_harness":"codex","source_row_index":222,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":10.56,"totalDurationS":50813,"totalInputTokens":46726983,"totalOutputTokens":263717,"total_cost_usd":10.56,"total_input_tokens":46726983,"total_output_tokens":263717,"total_runtime_s":50813,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.273103955698325,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c0402b6b5633cba67932","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8cf99679f407338ab7156c5f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.31023834901975544,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":566,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[282];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":566,"metadata":{"avgScore":0.31023834901975544,"avg_score":0.31023834901975544,"avg_score_percent":31.023834901975544,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.77,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":119921277,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":610291,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":152,"runtime":88490,"source_flags":[],"source_harness":"codex","source_row_index":282,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":26.77,"totalDurationS":88490,"totalInputTokens":119921277,"totalOutputTokens":610291,"total_cost_usd":26.77,"total_input_tokens":119921277,"total_output_tokens":610291,"total_runtime_s":88490,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.023834901975544,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fb6bd8f4d3a7d1d63ddb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-860f794638531334ebb2f962","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7562644213793285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1198,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[598];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1198,"metadata":{"avgScore":0.7562644213793285,"avg_score":0.7562644213793285,"avg_score_percent":75.62644213793284,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.71,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":359705378,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1633460,"passRate":0.4925373134328358,"pass_rate":0.4925373134328358,"pass_rate_percent":49.25373134328358,"passes":33,"runs":67,"runtime":69711,"source_flags":[],"source_harness":"codex","source_row_index":598,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":68.71,"totalDurationS":69711,"totalInputTokens":359705378,"totalOutputTokens":1633460,"total_cost_usd":68.71,"total_input_tokens":359705378,"total_output_tokens":1633460,"total_runtime_s":69711,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":75.62644213793284,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-287a0e893675c862e0ee","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8212a4cb1e34bd341ad4ad50","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.31023834901975544,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":94,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[46];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":94,"metadata":{"avgScore":0.31023834901975544,"avg_score":0.31023834901975544,"avg_score_percent":31.023834901975544,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.77,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":119921277,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":610291,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":152,"runtime":88490,"source_flags":[],"source_harness":"codex","source_row_index":46,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":26.77,"totalDurationS":88490,"totalInputTokens":119921277,"totalOutputTokens":610291,"total_cost_usd":26.77,"total_input_tokens":119921277,"total_output_tokens":610291,"total_runtime_s":88490,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.023834901975544,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-43051e8b551e20c88e43","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7867c2ff650149d8c18635c7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24564591975351033,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1050,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[524];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1050,"metadata":{"avgScore":0.24564591975351033,"avg_score":0.24564591975351033,"avg_score_percent":24.564591975351032,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":6.8,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":28407280,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":188827,"passRate":0.06,"pass_rate":0.06,"pass_rate_percent":6.0,"passes":3,"runs":50,"runtime":15568,"source_flags":[],"source_harness":"codex","source_row_index":524,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":6.8,"totalDurationS":15568,"totalInputTokens":28407280,"totalOutputTokens":188827,"total_cost_usd":6.8,"total_input_tokens":28407280,"total_output_tokens":188827,"total_runtime_s":15568,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.564591975351032,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1c9da1fb345d5bc7e0e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6ec386e220ad565db5f7c340","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5037352082410917,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":492,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[245];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":492,"metadata":{"avgScore":0.5037352082410917,"avg_score":0.5037352082410917,"avg_score_percent":50.37352082410918,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":390.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2499272937,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":6752715,"passRate":0.2894736842105263,"pass_rate":0.2894736842105263,"pass_rate_percent":28.947368421052634,"passes":44,"runs":149,"runtime":239192,"source_flags":[],"source_harness":"codex","source_row_index":245,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":390.61,"totalDurationS":239192,"totalInputTokens":2499272937,"totalOutputTokens":6752715,"total_cost_usd":390.61,"total_input_tokens":2499272937,"total_output_tokens":6752715,"total_runtime_s":239192,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.37352082410918,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-117635df65e5881d1a2e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6e77139565a32ea7a46a3043","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6736742823508564,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1224,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[611];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1224,"metadata":{"avgScore":0.6736742823508564,"avg_score":0.6736742823508564,"avg_score_percent":67.36742823508564,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":228811254,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1065059,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":67,"runtime":60584,"source_flags":[],"source_harness":"codex","source_row_index":611,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":46.18,"totalDurationS":60584,"totalInputTokens":228811254,"totalOutputTokens":1065059,"total_cost_usd":46.18,"total_input_tokens":228811254,"total_output_tokens":1065059,"total_runtime_s":60584,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.36742823508564,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5819b6238d0548d1d801","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6d1e6f3cc5f545001288da0b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.743547028673177,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":366,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[182];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":366,"metadata":{"avgScore":0.743547028673177,"avg_score":0.743547028673177,"avg_score_percent":74.3547028673177,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":541671893,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2191134,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":67,"runtime":78518,"source_flags":[],"source_harness":"codex","source_row_index":182,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.67,"totalDurationS":78518,"totalInputTokens":541671893,"totalOutputTokens":2191134,"total_cost_usd":95.67,"total_input_tokens":541671893,"total_output_tokens":2191134,"total_runtime_s":78518,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":74.3547028673177,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8bfe459457fdd3961606","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-64415c51f00e0f41e1f20621","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3280611890351934,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":930,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[464];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":930,"metadata":{"avgScore":0.3280611890351934,"avg_score":0.3280611890351934,"avg_score_percent":32.80611890351934,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":102309910,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":562218,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":141,"runtime":77880,"source_flags":[],"source_harness":"codex","source_row_index":464,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":22.46,"totalDurationS":77880,"totalInputTokens":102309910,"totalOutputTokens":562218,"total_cost_usd":22.46,"total_input_tokens":102309910,"total_output_tokens":562218,"total_runtime_s":77880,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.80611890351934,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-45448765d7fbdcb3b782","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-632d0d6ba52895a590c55f8a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3850041500197,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":546,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[272];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":546,"metadata":{"avgScore":0.3850041500197,"avg_score":0.3850041500197,"avg_score_percent":38.50041500197,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":57,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":287164350,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1082713,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":151697,"source_flags":[],"source_harness":"codex","source_row_index":272,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":57,"totalDurationS":151697,"totalInputTokens":287164350,"totalOutputTokens":1082713,"total_cost_usd":57,"total_input_tokens":287164350,"total_output_tokens":1082713,"total_runtime_s":151697,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.50041500197,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bd7382209d16651e210c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5e0760df54c3e20eae869b06","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1855642458507263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":640,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[319];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":640,"metadata":{"avgScore":0.1855642458507263,"avg_score":0.1855642458507263,"avg_score_percent":18.55642458507263,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.28,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":63964976,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":85737,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":11476,"source_flags":[],"source_harness":"codex","source_row_index":319,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":12.28,"totalDurationS":11476,"totalInputTokens":63964976,"totalOutputTokens":85737,"total_cost_usd":12.28,"total_input_tokens":63964976,"total_output_tokens":85737,"total_runtime_s":11476,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.55642458507263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-996aee14014f9de8b551","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5d6342079c87b1c6e2eaf27c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.07960916226247187,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1160,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[579];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1160,"metadata":{"avgScore":0.07960916226247187,"avg_score":0.07960916226247187,"avg_score_percent":7.960916226247187,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":39.73,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":253840806,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":675388,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":59005,"source_flags":[],"source_harness":"codex","source_row_index":579,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":39.73,"totalDurationS":59005,"totalInputTokens":253840806,"totalOutputTokens":675388,"total_cost_usd":39.73,"total_input_tokens":253840806,"total_output_tokens":675388,"total_runtime_s":59005,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.960916226247187,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-953a31877af88b02b5e8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5be0b289b950bd0740611d62","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10395747737317527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":300,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[149];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":300,"metadata":{"avgScore":0.10395747737317527,"avg_score":0.10395747737317527,"avg_score_percent":10.395747737317528,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":23.91,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":132807946,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":319379,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":56504,"source_flags":[],"source_harness":"codex","source_row_index":149,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":23.91,"totalDurationS":56504,"totalInputTokens":132807946,"totalOutputTokens":319379,"total_cost_usd":23.91,"total_input_tokens":132807946,"total_output_tokens":319379,"total_runtime_s":56504,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.395747737317528,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-366edfb65230d81b120f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5b5f954bc6def8301500371c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.41712583069964326,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":148,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[73];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":148,"metadata":{"avgScore":0.41712583069964326,"avg_score":0.41712583069964326,"avg_score_percent":41.712583069964325,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":351830472,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1354167,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":15,"runs":55,"runtime":81285,"source_flags":[],"source_harness":"codex","source_row_index":73,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":61.13,"totalDurationS":81285,"totalInputTokens":351830472,"totalOutputTokens":1354167,"total_cost_usd":61.13,"total_input_tokens":351830472,"total_output_tokens":1354167,"total_runtime_s":81285,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.712583069964325,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92090848b57d6195ed6a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-577ef54bde84ac7ad2d8272a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6092947556539312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":416,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[207];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":416,"metadata":{"avgScore":0.6092947556539312,"avg_score":0.6092947556539312,"avg_score_percent":60.92947556539312,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":19.45,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":90618587,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":452393,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":67,"runtime":67571,"source_flags":[],"source_harness":"codex","source_row_index":207,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":19.45,"totalDurationS":67571,"totalInputTokens":90618587,"totalOutputTokens":452393,"total_cost_usd":19.45,"total_input_tokens":90618587,"total_output_tokens":452393,"total_runtime_s":67571,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.92947556539312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ad123aa8b61577f2a6b6","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-56f53630077d22f8d7c4bb72","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24891449790392856,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":208,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[103];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":208,"metadata":{"avgScore":0.24891449790392856,"avg_score":0.24891449790392856,"avg_score_percent":24.891449790392855,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":9.01,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":36303840,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":207513,"passRate":0.07272727272727272,"pass_rate":0.07272727272727272,"pass_rate_percent":7.2727272727272725,"passes":4,"runs":55,"runtime":19943,"source_flags":[],"source_harness":"codex","source_row_index":103,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":9.01,"totalDurationS":19943,"totalInputTokens":36303840,"totalOutputTokens":207513,"total_cost_usd":9.01,"total_input_tokens":36303840,"total_output_tokens":207513,"total_runtime_s":19943,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.891449790392855,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6b12fae4c363bf8cf7e4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-531eeee3df6d32c488e347c5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5331379784041669,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1320,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[659];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1320,"metadata":{"avgScore":0.5331379784041669,"avg_score":0.5331379784041669,"avg_score_percent":53.31379784041669,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":371.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2398634685,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":6446026,"passRate":0.3120567375886525,"pass_rate":0.3120567375886525,"pass_rate_percent":31.20567375886525,"passes":44,"runs":141,"runtime":199793,"source_flags":[],"source_harness":"codex","source_row_index":659,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":371.53,"totalDurationS":199793,"totalInputTokens":2398634685,"totalOutputTokens":6446026,"total_cost_usd":371.53,"total_input_tokens":2398634685,"total_output_tokens":6446026,"total_runtime_s":199793,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.31379784041669,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-12ba7d2a58ee10cd6050","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-52e39d44dc1451073f76396f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.47755529230158056,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":742,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[370];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":742,"metadata":{"avgScore":0.47755529230158056,"avg_score":0.47755529230158056,"avg_score_percent":47.755529230158054,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":72.92,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":413704161,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1707402,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":105,"runtime":60549,"source_flags":[],"source_harness":"codex","source_row_index":370,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":72.92,"totalDurationS":60549,"totalInputTokens":413704161,"totalOutputTokens":1707402,"total_cost_usd":72.92,"total_input_tokens":413704161,"total_output_tokens":1707402,"total_runtime_s":60549,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.755529230158054,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-01459e5d0a1cba566e6f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4db23db1c55dd48d8f697b71","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1268451543325831,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":656,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[327];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":656,"metadata":{"avgScore":0.1268451543325831,"avg_score":0.1268451543325831,"avg_score_percent":12.684515433258309,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":19.08,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":100638252,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":306689,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":8,"runtime":39399,"source_flags":[],"source_harness":"codex","source_row_index":327,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":8,"tokens":null,"totalCostUsd":19.08,"totalDurationS":39399,"totalInputTokens":100638252,"totalOutputTokens":306689,"total_cost_usd":19.08,"total_input_tokens":100638252,"total_output_tokens":306689,"total_runtime_s":39399,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.684515433258309,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1df7d077505b773b890e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-480f179605d41209a03a67d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4356378222921257,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":970,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[484];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":970,"metadata":{"avgScore":0.4356378222921257,"avg_score":0.4356378222921257,"avg_score_percent":43.56378222921257,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":33.54,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":174023797,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":768309,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":14,"runs":50,"runtime":26559,"source_flags":[],"source_harness":"codex","source_row_index":484,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":33.54,"totalDurationS":26559,"totalInputTokens":174023797,"totalOutputTokens":768309,"total_cost_usd":33.54,"total_input_tokens":174023797,"total_output_tokens":768309,"total_runtime_s":26559,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.56378222921257,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c6caa907ae4d6a0230f2","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-45a0d469978fec9bb0616caf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.48273103955698327,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1282,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[640];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1282,"metadata":{"avgScore":0.48273103955698327,"avg_score":0.48273103955698327,"avg_score_percent":48.273103955698325,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":10.56,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":46726983,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":263717,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":67,"runtime":50813,"source_flags":[],"source_harness":"codex","source_row_index":640,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":10.56,"totalDurationS":50813,"totalInputTokens":46726983,"totalOutputTokens":263717,"total_cost_usd":10.56,"total_input_tokens":46726983,"total_output_tokens":263717,"total_runtime_s":50813,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.273103955698325,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3dedb09f5ad45c1f9bf5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3c8a1944ced12ef65b2c38a0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5238739062483173,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":716,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[357];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":716,"metadata":{"avgScore":0.5238739062483173,"avg_score":0.5238739062483173,"avg_score_percent":52.38739062483173,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.38,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1605477710,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":4271202,"passRate":0.2857142857142857,"pass_rate":0.2857142857142857,"pass_rate_percent":28.57142857142857,"passes":30,"runs":105,"runtime":127293,"source_flags":[],"source_harness":"codex","source_row_index":357,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":246.38,"totalDurationS":127293,"totalInputTokens":1605477710,"totalOutputTokens":4271202,"total_cost_usd":246.38,"total_input_tokens":1605477710,"total_output_tokens":4271202,"total_runtime_s":127293,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.38739062483173,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-179bd864572ea69c214e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-388bd27514db2d168f4538ce","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6092947556539312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1252,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[625];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1252,"metadata":{"avgScore":0.6092947556539312,"avg_score":0.6092947556539312,"avg_score_percent":60.92947556539312,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":19.45,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":90618587,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":452393,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":67,"runtime":67571,"source_flags":[],"source_harness":"codex","source_row_index":625,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":19.45,"totalDurationS":67571,"totalInputTokens":90618587,"totalOutputTokens":452393,"total_cost_usd":19.45,"total_input_tokens":90618587,"total_output_tokens":452393,"total_runtime_s":67571,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.92947556539312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8f0df7aba62d6f028818","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-35591a8371b1502f86442299","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3850041500197,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":74,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[36];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":74,"metadata":{"avgScore":0.3850041500197,"avg_score":0.3850041500197,"avg_score_percent":38.50041500197,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":57,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":287164350,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":1082713,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":151697,"source_flags":[],"source_harness":"codex","source_row_index":36,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":57,"totalDurationS":151697,"totalInputTokens":287164350,"totalOutputTokens":1082713,"total_cost_usd":57,"total_input_tokens":287164350,"total_output_tokens":1082713,"total_runtime_s":151697,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.50041500197,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-649b4064353f1e2f8058","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-35444cf18c62af42d87a1d26","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.41185751264390447,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":780,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[389];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":780,"metadata":{"avgScore":0.41185751264390447,"avg_score":0.41185751264390447,"avg_score_percent":41.18575126439045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":25.06,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":127297515,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":721721,"passRate":0.18095238095238095,"pass_rate":0.18095238095238095,"pass_rate_percent":18.095238095238095,"passes":19,"runs":105,"runtime":48741,"source_flags":[],"source_harness":"codex","source_row_index":389,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":25.06,"totalDurationS":48741,"totalInputTokens":127297515,"totalOutputTokens":721721,"total_cost_usd":25.06,"total_input_tokens":127297515,"total_output_tokens":721721,"total_runtime_s":48741,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.18575126439045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6126a32c8ca662af22cd","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3365a605bff0b0db17c5902b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.08890663199918114,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":310,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[154];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":310,"metadata":{"avgScore":0.08890663199918114,"avg_score":0.08890663199918114,"avg_score_percent":8.890663199918114,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.62,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":42591540,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":162532,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":22384,"source_flags":[],"source_harness":"codex","source_row_index":154,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":8.62,"totalDurationS":22384,"totalInputTokens":42591540,"totalOutputTokens":162532,"total_cost_usd":8.62,"total_input_tokens":42591540,"total_output_tokens":162532,"total_runtime_s":22384,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":8.890663199918114,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-79daad14bf874cc8e605","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-32bf3da9fbc76d974efba0e4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.08178194518550498,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":670,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[334];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":670,"metadata":{"avgScore":0.08178194518550498,"avg_score":0.08178194518550498,"avg_score_percent":8.178194518550498,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4.31,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":17611367,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":48073,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":10610,"source_flags":[],"source_harness":"codex","source_row_index":334,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":4.31,"totalDurationS":10610,"totalInputTokens":17611367,"totalOutputTokens":48073,"total_cost_usd":4.31,"total_input_tokens":17611367,"total_output_tokens":48073,"total_runtime_s":10610,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":8.178194518550498,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-abd315c967e478129ce7","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-30a681b01fe903c0a575e94f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4433841810461951,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":966,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[482];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":966,"metadata":{"avgScore":0.4433841810461951,"avg_score":0.4433841810461951,"avg_score_percent":44.33841810461951,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.07,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":585468127,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2068701,"passRate":0.3,"pass_rate":0.3,"pass_rate_percent":30.0,"passes":15,"runs":50,"runtime":47507,"source_flags":[],"source_harness":"codex","source_row_index":482,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":95.07,"totalDurationS":47507,"totalInputTokens":585468127,"totalOutputTokens":2068701,"total_cost_usd":95.07,"total_input_tokens":585468127,"total_output_tokens":2068701,"total_runtime_s":47507,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.33841810461951,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-078ad389dde2076b2b9a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2116bd74dc3c209356b1edfc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4789918638067592,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1342,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[670];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1342,"metadata":{"avgScore":0.4789918638067592,"avg_score":0.4789918638067592,"avg_score_percent":47.899186380675914,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":116.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":643244149,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2416086,"passRate":0.2624113475177305,"pass_rate":0.2624113475177305,"pass_rate_percent":26.24113475177305,"passes":37,"runs":141,"runtime":142348,"source_flags":[],"source_harness":"codex","source_row_index":670,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":116.46,"totalDurationS":142348,"totalInputTokens":643244149,"totalOutputTokens":2416086,"total_cost_usd":116.46,"total_input_tokens":643244149,"total_output_tokens":2416086,"total_runtime_s":142348,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.899186380675914,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0927426632670ea9928f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1308eefacfc0293fb0354bff","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.743547028673177,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1202,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[600];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1202,"metadata":{"avgScore":0.743547028673177,"avg_score":0.743547028673177,"avg_score_percent":74.3547028673177,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":541671893,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2191134,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":67,"runtime":78518,"source_flags":[],"source_harness":"codex","source_row_index":600,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.67,"totalDurationS":78518,"totalInputTokens":541671893,"totalOutputTokens":2191134,"total_cost_usd":95.67,"total_input_tokens":541671893,"total_output_tokens":2191134,"total_runtime_s":78518,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":74.3547028673177,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-212db766d0a1f3c4d915","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-00d79827380c1e42a28d8440","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1496761918512072,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5efbf9a0d7e9ae1e7d8b59c90d5812b523191b4","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":260,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[129];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":260,"metadata":{"avgScore":0.1496761918512072,"avg_score":0.1496761918512072,"avg_score_percent":14.96761918512072,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":197,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1356495846,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Luna","output_tokens":2624201,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":36,"runtime":113686,"source_flags":[],"source_harness":"codex","source_row_index":129,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":197,"totalDurationS":113686,"totalInputTokens":1356495846,"totalOutputTokens":2624201,"total_cost_usd":197,"total_input_tokens":1356495846,"total_output_tokens":2624201,"total_runtime_s":113686,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.96761918512072,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ee18fa82eff20738035f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fe28594badc57e44f221fe9f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3582089552238806,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":387,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[193];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":387,"metadata":{"avgScore":0.6736742823508564,"avg_score":0.6736742823508564,"avg_score_percent":67.36742823508564,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":228811254,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1065059,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":67,"runtime":60584,"source_flags":[],"source_harness":"codex","source_row_index":193,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":46.18,"totalDurationS":60584,"totalInputTokens":228811254,"totalOutputTokens":1065059,"total_cost_usd":46.18,"total_input_tokens":228811254,"total_output_tokens":1065059,"total_runtime_s":60584,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.82089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2904ee1ab1c7bcd36bde","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fa8035be3a321e36b36250a7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.43283582089552236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":365,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[182];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":365,"metadata":{"avgScore":0.743547028673177,"avg_score":0.743547028673177,"avg_score_percent":74.3547028673177,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":541671893,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2191134,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":67,"runtime":78518,"source_flags":[],"source_harness":"codex","source_row_index":182,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.67,"totalDurationS":78518,"totalInputTokens":541671893,"totalOutputTokens":2191134,"total_cost_usd":95.67,"total_input_tokens":541671893,"total_output_tokens":2191134,"total_runtime_s":78518,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.28358208955223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8c9a617f61ce652d5b0d","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f758ca08b30d50b57c728a4c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3120567375886525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1319,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[659];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1319,"metadata":{"avgScore":0.5331379784041669,"avg_score":0.5331379784041669,"avg_score_percent":53.31379784041669,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":371.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2398634685,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":6446026,"passRate":0.3120567375886525,"pass_rate":0.3120567375886525,"pass_rate_percent":31.20567375886525,"passes":44,"runs":141,"runtime":199793,"source_flags":[],"source_harness":"codex","source_row_index":659,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":371.53,"totalDurationS":199793,"totalInputTokens":2398634685,"totalOutputTokens":6446026,"total_cost_usd":371.53,"total_input_tokens":2398634685,"total_output_tokens":6446026,"total_runtime_s":199793,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.20567375886525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3d19e5bdce7af5dbd65d","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f68fd72068c7b893e3751ec8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2857142857142857,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":715,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[357];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":715,"metadata":{"avgScore":0.5238739062483173,"avg_score":0.5238739062483173,"avg_score_percent":52.38739062483173,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.38,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1605477710,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":4271202,"passRate":0.2857142857142857,"pass_rate":0.2857142857142857,"pass_rate_percent":28.57142857142857,"passes":30,"runs":105,"runtime":127293,"source_flags":[],"source_harness":"codex","source_row_index":357,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":246.38,"totalDurationS":127293,"totalInputTokens":1605477710,"totalOutputTokens":4271202,"total_cost_usd":246.38,"total_input_tokens":1605477710,"total_output_tokens":4271202,"total_runtime_s":127293,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.57142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2dbd637e2e70fb1be9e9","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f4a51ce97d490c8d62e7cd72","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1276595744680851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":929,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[464];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":929,"metadata":{"avgScore":0.3280611890351934,"avg_score":0.3280611890351934,"avg_score_percent":32.80611890351934,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":102309910,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":562218,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":141,"runtime":77880,"source_flags":[],"source_harness":"codex","source_row_index":464,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":22.46,"totalDurationS":77880,"totalInputTokens":102309910,"totalOutputTokens":562218,"total_cost_usd":22.46,"total_input_tokens":102309910,"total_output_tokens":562218,"total_runtime_s":77880,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.76595744680851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7b394687853ae96031ba","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f41d071e980ef4c2e045f4ed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":143,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[71];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":143,"metadata":{"avgScore":0.4253164625672212,"avg_score":0.4253164625672212,"avg_score_percent":42.53164625672212,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.34,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":652436878,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2206867,"passRate":0.2909090909090909,"pass_rate":0.2909090909090909,"pass_rate_percent":29.09090909090909,"passes":16,"runs":54,"runtime":71516,"source_flags":[],"source_harness":"codex","source_row_index":71,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":107.34,"totalDurationS":71516,"totalInputTokens":652436878,"totalOutputTokens":2206867,"total_cost_usd":107.34,"total_input_tokens":652436878,"total_output_tokens":2206867,"total_runtime_s":71516,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.09090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5e68bc75bbaf0de39c3e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-effc46595853b26de7618c2d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.12,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1035,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[517];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1035,"metadata":{"avgScore":0.2963635563049718,"avg_score":0.2963635563049718,"avg_score_percent":29.636355630497178,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":11.82,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":51495533,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":311061,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":50,"runtime":27012,"source_flags":[],"source_harness":"codex","source_row_index":517,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":11.82,"totalDurationS":27012,"totalInputTokens":51495533,"totalOutputTokens":311061,"total_cost_usd":11.82,"total_input_tokens":51495533,"total_output_tokens":311061,"total_runtime_s":27012,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e7c7735e5cc02c280ccf","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ed73792a27ab9d114e0535d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.11842105263157894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":565,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[282];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":565,"metadata":{"avgScore":0.31023834901975544,"avg_score":0.31023834901975544,"avg_score_percent":31.023834901975544,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.77,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":119921277,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":610291,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":152,"runtime":88490,"source_flags":[],"source_harness":"codex","source_row_index":282,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":26.77,"totalDurationS":88490,"totalInputTokens":119921277,"totalOutputTokens":610291,"total_cost_usd":26.77,"total_input_tokens":119921277,"total_output_tokens":610291,"total_runtime_s":88490,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.842105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f8307060d5855c01292c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e8ab76b9fa7b0bb0b2125b5e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22388059701492538,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1281,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[640];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1281,"metadata":{"avgScore":0.48273103955698327,"avg_score":0.48273103955698327,"avg_score_percent":48.273103955698325,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":10.56,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":46726983,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":263717,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":67,"runtime":50813,"source_flags":[],"source_harness":"codex","source_row_index":640,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":10.56,"totalDurationS":50813,"totalInputTokens":46726983,"totalOutputTokens":263717,"total_cost_usd":10.56,"total_input_tokens":46726983,"total_output_tokens":263717,"total_runtime_s":50813,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.388059701492537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cc9002c33f78fe5ad2f5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-dfc9b9af3653b0073130ecac","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18095238095238095,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":779,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[389];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":779,"metadata":{"avgScore":0.41185751264390447,"avg_score":0.41185751264390447,"avg_score_percent":41.18575126439045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":25.06,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":127297515,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":721721,"passRate":0.18095238095238095,"pass_rate":0.18095238095238095,"pass_rate_percent":18.095238095238095,"passes":19,"runs":105,"runtime":48741,"source_flags":[],"source_harness":"codex","source_row_index":389,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":25.06,"totalDurationS":48741,"totalInputTokens":127297515,"totalOutputTokens":721721,"total_cost_usd":25.06,"total_input_tokens":127297515,"total_output_tokens":721721,"total_runtime_s":48741,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.095238095238095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd74f328485f2fc6ff18","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-db04de1c6450c0db49dffa3f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1157,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[578];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1157,"metadata":{"avgScore":0.07978474570120461,"avg_score":0.07978474570120461,"avg_score_percent":7.978474570120461,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":14.4,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":84561224,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":266098,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":48999,"source_flags":[],"source_harness":"codex","source_row_index":578,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":14.4,"totalDurationS":48999,"totalInputTokens":84561224,"totalOutputTokens":266098,"total_cost_usd":14.4,"total_input_tokens":84561224,"total_output_tokens":266098,"total_runtime_s":48999,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3afb6e7733e08ba75677","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d8719fb0c243e2f53834302d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24342105263157895,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":505,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[252];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":505,"metadata":{"avgScore":0.46128766882143674,"avg_score":0.46128766882143674,"avg_score_percent":46.12876688214367,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.52,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":699354083,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2559522,"passRate":0.24342105263157895,"pass_rate":0.24342105263157895,"pass_rate_percent":24.342105263157894,"passes":37,"runs":152,"runtime":165192,"source_flags":[],"source_harness":"codex","source_row_index":252,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":127.52,"totalDurationS":165192,"totalInputTokens":699354083,"totalOutputTokens":2559522,"total_cost_usd":127.52,"total_input_tokens":699354083,"total_output_tokens":2559522,"total_runtime_s":165192,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.342105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8021c2259759816474c8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d69e558d871c60b09c7d1743","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":309,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[154];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":309,"metadata":{"avgScore":0.08890663199918114,"avg_score":0.08890663199918114,"avg_score_percent":8.890663199918114,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.62,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":42591540,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":162532,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":22384,"source_flags":[],"source_harness":"codex","source_row_index":154,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":8.62,"totalDurationS":22384,"totalInputTokens":42591540,"totalOutputTokens":162532,"total_cost_usd":8.62,"total_input_tokens":42591540,"total_output_tokens":162532,"total_runtime_s":22384,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5987c2293fcb8eb6a767","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d19648c59c7731f016f60ebe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.07272727272727272,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":207,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[103];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":207,"metadata":{"avgScore":0.24891449790392856,"avg_score":0.24891449790392856,"avg_score_percent":24.891449790392855,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":9.01,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":36303840,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":207513,"passRate":0.07272727272727272,"pass_rate":0.07272727272727272,"pass_rate_percent":7.2727272727272725,"passes":4,"runs":55,"runtime":19943,"source_flags":[],"source_harness":"codex","source_row_index":103,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":9.01,"totalDurationS":19943,"totalInputTokens":36303840,"totalOutputTokens":207513,"total_cost_usd":9.01,"total_input_tokens":36303840,"total_output_tokens":207513,"total_runtime_s":19943,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.2727272727272725,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-05cd9939afecad18b13b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c5429a3fb99a4437e375a73d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":639,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[319];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":639,"metadata":{"avgScore":0.1855642458507263,"avg_score":0.1855642458507263,"avg_score_percent":18.55642458507263,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.28,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":63964976,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":85737,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":11476,"source_flags":[],"source_harness":"codex","source_row_index":319,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":12.28,"totalDurationS":11476,"totalInputTokens":63964976,"totalOutputTokens":85737,"total_cost_usd":12.28,"total_input_tokens":63964976,"total_output_tokens":85737,"total_runtime_s":11476,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e6809e5433b7bca9f9c9","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c402160d49e5a711ffee920b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1159,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[579];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1159,"metadata":{"avgScore":0.07960916226247187,"avg_score":0.07960916226247187,"avg_score_percent":7.960916226247187,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":39.73,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":253840806,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":675388,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":59005,"source_flags":[],"source_harness":"codex","source_row_index":579,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":39.73,"totalDurationS":59005,"totalInputTokens":253840806,"totalOutputTokens":675388,"total_cost_usd":39.73,"total_input_tokens":253840806,"total_output_tokens":675388,"total_runtime_s":59005,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e0fdcc5198821067681e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c3ad68c1b5bb805c071a9167","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3120567375886525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":847,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[423];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":847,"metadata":{"avgScore":0.5331379784041669,"avg_score":0.5331379784041669,"avg_score_percent":53.31379784041669,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":371.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2398634685,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":6446026,"passRate":0.3120567375886525,"pass_rate":0.3120567375886525,"pass_rate_percent":31.20567375886525,"passes":44,"runs":141,"runtime":199793,"source_flags":[],"source_harness":"codex","source_row_index":423,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":371.53,"totalDurationS":199793,"totalInputTokens":2398634685,"totalOutputTokens":6446026,"total_cost_usd":371.53,"total_input_tokens":2398634685,"total_output_tokens":6446026,"total_runtime_s":199793,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.20567375886525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-893f6096cb727786a60c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bfdf29433b8d8cda37ea40a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24761904761904763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":741,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[370];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":741,"metadata":{"avgScore":0.47755529230158056,"avg_score":0.47755529230158056,"avg_score_percent":47.755529230158054,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":72.92,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":413704161,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1707402,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":105,"runtime":60549,"source_flags":[],"source_harness":"codex","source_row_index":370,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":72.92,"totalDurationS":60549,"totalInputTokens":413704161,"totalOutputTokens":1707402,"total_cost_usd":72.92,"total_input_tokens":413704161,"total_output_tokens":1707402,"total_runtime_s":60549,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.761904761904763,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-076704d122953a4aa0da","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-be9c83647b5cadc91c1f5a51","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":289,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[144];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":289,"metadata":{"avgScore":0.10970741091203197,"avg_score":0.10970741091203197,"avg_score_percent":10.970741091203196,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":109.66,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":742917153,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1343720,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":108705,"source_flags":[],"source_harness":"codex","source_row_index":144,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":109.66,"totalDurationS":108705,"totalInputTokens":742917153,"totalOutputTokens":1343720,"total_cost_usd":109.66,"total_input_tokens":742917153,"total_output_tokens":1343720,"total_runtime_s":108705,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-856165e35a6479739663","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bb4a1928a152a58ea1519b25","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1276595744680851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1401,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[700];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1401,"metadata":{"avgScore":0.3280611890351934,"avg_score":0.3280611890351934,"avg_score_percent":32.80611890351934,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":102309910,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":562218,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":141,"runtime":77880,"source_flags":[],"source_harness":"codex","source_row_index":700,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":22.46,"totalDurationS":77880,"totalInputTokens":102309910,"totalOutputTokens":562218,"total_cost_usd":22.46,"total_input_tokens":102309910,"total_output_tokens":562218,"total_runtime_s":77880,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.76595744680851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5897bf94c421c25938ea","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-adc6d0b2c403d338dbba852e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2727272727272727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":135,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[67];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":135,"metadata":{"avgScore":0.4372134877427408,"avg_score":0.4372134877427408,"avg_score_percent":43.721348774274084,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.31,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":202023719,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":828678,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":15,"runs":55,"runtime":37502,"source_flags":[],"source_harness":"codex","source_row_index":67,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":38.31,"totalDurationS":37502,"totalInputTokens":202023719,"totalOutputTokens":828678,"total_cost_usd":38.31,"total_input_tokens":202023719,"total_output_tokens":828678,"total_runtime_s":37502,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c173247d5437fdd181b3","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a803ddac5a96ea86071d03fa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2624113475177305,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":869,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[434];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":869,"metadata":{"avgScore":0.4789918638067592,"avg_score":0.4789918638067592,"avg_score_percent":47.899186380675914,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":116.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":643244149,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2416086,"passRate":0.2624113475177305,"pass_rate":0.2624113475177305,"pass_rate_percent":26.24113475177305,"passes":37,"runs":141,"runtime":142348,"source_flags":[],"source_harness":"codex","source_row_index":434,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":116.46,"totalDurationS":142348,"totalInputTokens":643244149,"totalOutputTokens":2416086,"total_cost_usd":116.46,"total_input_tokens":643244149,"total_output_tokens":2416086,"total_runtime_s":142348,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.24113475177305,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-37c391b63487615de2bf","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a6108154d2cc500d9a048903","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3262411347517731,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1327,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[663];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1327,"metadata":{"avgScore":0.5206445104262789,"avg_score":0.5206445104262789,"avg_score_percent":52.064451042627894,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":212.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1305756331,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":3877107,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":46,"runs":141,"runtime":189252,"source_flags":[],"source_harness":"codex","source_row_index":663,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":212.65,"totalDurationS":189252,"totalInputTokens":1305756331,"totalOutputTokens":3877107,"total_cost_usd":212.65,"total_input_tokens":1305756331,"total_output_tokens":3877107,"total_runtime_s":189252,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.62411347517731,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-349912c1f8eaa868b1bb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a5ec43e616951c2c22ba389f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24342105263157895,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":33,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[16];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":33,"metadata":{"avgScore":0.46128766882143674,"avg_score":0.46128766882143674,"avg_score_percent":46.12876688214367,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.52,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":699354083,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2559522,"passRate":0.24342105263157895,"pass_rate":0.24342105263157895,"pass_rate_percent":24.342105263157894,"passes":37,"runs":152,"runtime":165192,"source_flags":[],"source_harness":"codex","source_row_index":16,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":127.52,"totalDurationS":165192,"totalInputTokens":699354083,"totalOutputTokens":2559522,"total_cost_usd":127.52,"total_input_tokens":699354083,"total_output_tokens":2559522,"total_runtime_s":165192,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.342105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c60d9e35c86cc143155b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-9f6f4ecd7d4744a1c5424e11","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":969,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[484];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":969,"metadata":{"avgScore":0.4356378222921257,"avg_score":0.4356378222921257,"avg_score_percent":43.56378222921257,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":33.54,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":174023797,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":768309,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":14,"runs":50,"runtime":26559,"source_flags":[],"source_harness":"codex","source_row_index":484,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":33.54,"totalDurationS":26559,"totalInputTokens":174023797,"totalOutputTokens":768309,"total_cost_usd":33.54,"total_input_tokens":174023797,"total_output_tokens":768309,"total_runtime_s":26559,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ee80201f6500f09e51f9","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-9e8efe3e531a8187bcc4bd84","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29523809523809524,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":725,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[362];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":725,"metadata":{"avgScore":0.507344624718011,"avg_score":0.507344624718011,"avg_score_percent":50.7344624718011,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":143.63,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":910676386,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2836040,"passRate":0.29523809523809524,"pass_rate":0.29523809523809524,"pass_rate_percent":29.523809523809526,"passes":31,"runs":105,"runtime":89570,"source_flags":[],"source_harness":"codex","source_row_index":362,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":143.63,"totalDurationS":89570,"totalInputTokens":910676386,"totalOutputTokens":2836040,"total_cost_usd":143.63,"total_input_tokens":910676386,"total_output_tokens":2836040,"total_runtime_s":89570,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.523809523809526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0d48c45b6652e0ccab66","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-97518880af13d2cb8450a090","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.11842105263157894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":93,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[46];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":93,"metadata":{"avgScore":0.31023834901975544,"avg_score":0.31023834901975544,"avg_score_percent":31.023834901975544,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.77,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":119921277,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":610291,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":152,"runtime":88490,"source_flags":[],"source_harness":"codex","source_row_index":46,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":26.77,"totalDurationS":88490,"totalInputTokens":119921277,"totalOutputTokens":610291,"total_cost_usd":26.77,"total_input_tokens":119921277,"total_output_tokens":610291,"total_runtime_s":88490,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.842105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a3ffc2833b9aadc6694","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8ec819d8e58851167c118606","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2624113475177305,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1341,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[670];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1341,"metadata":{"avgScore":0.4789918638067592,"avg_score":0.4789918638067592,"avg_score_percent":47.899186380675914,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":116.46,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":643244149,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2416086,"passRate":0.2624113475177305,"pass_rate":0.2624113475177305,"pass_rate_percent":26.24113475177305,"passes":37,"runs":141,"runtime":142348,"source_flags":[],"source_harness":"codex","source_row_index":670,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":116.46,"totalDurationS":142348,"totalInputTokens":643244149,"totalOutputTokens":2416086,"total_cost_usd":116.46,"total_input_tokens":643244149,"total_output_tokens":2416086,"total_runtime_s":142348,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.24113475177305,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4a427d762febaf9137c9","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8dc731de6e3aae718d4ce351","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":287,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[143];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":287,"metadata":{"avgScore":0.10993413212460641,"avg_score":0.10993413212460641,"avg_score_percent":10.993413212460641,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.64,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":284905745,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":772005,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":75483,"source_flags":[],"source_harness":"codex","source_row_index":143,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":46.64,"totalDurationS":75483,"totalInputTokens":284905745,"totalOutputTokens":772005,"total_cost_usd":46.64,"total_input_tokens":284905745,"total_output_tokens":772005,"total_runtime_s":75483,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e29e21c9d091e5bba8be","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8a525852e7bdedf37952a121","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":259,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[129];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":259,"metadata":{"avgScore":0.1496761918512072,"avg_score":0.1496761918512072,"avg_score_percent":14.96761918512072,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":197,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1356495846,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2624201,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":36,"runtime":113686,"source_flags":[],"source_harness":"codex","source_row_index":129,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":197,"totalDurationS":113686,"totalInputTokens":1356495846,"totalOutputTokens":2624201,"total_cost_usd":197,"total_input_tokens":1356495846,"total_output_tokens":2624201,"total_runtime_s":113686,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-196a6d61b534dca2fad5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-802963b4e4f9f455b401ee8e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4925373134328358,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":361,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[180];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":361,"metadata":{"avgScore":0.7562644213793285,"avg_score":0.7562644213793285,"avg_score_percent":75.62644213793284,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.71,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":359705378,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1633460,"passRate":0.4925373134328358,"pass_rate":0.4925373134328358,"pass_rate_percent":49.25373134328358,"passes":33,"runs":67,"runtime":69711,"source_flags":[],"source_harness":"codex","source_row_index":180,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":68.71,"totalDurationS":69711,"totalInputTokens":359705378,"totalOutputTokens":1633460,"total_cost_usd":68.71,"total_input_tokens":359705378,"total_output_tokens":1633460,"total_runtime_s":69711,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.25373134328358,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fe657a8c4ea46c96e04a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7bca23ca7529fa12cd13483f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2894736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":491,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[245];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":491,"metadata":{"avgScore":0.5037352082410917,"avg_score":0.5037352082410917,"avg_score_percent":50.37352082410918,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":390.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2499272937,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":6752715,"passRate":0.2894736842105263,"pass_rate":0.2894736842105263,"pass_rate_percent":28.947368421052634,"passes":44,"runs":149,"runtime":239192,"source_flags":[],"source_harness":"codex","source_row_index":245,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":390.61,"totalDurationS":239192,"totalInputTokens":2499272937,"totalOutputTokens":6752715,"total_cost_usd":390.61,"total_input_tokens":2499272937,"total_output_tokens":6752715,"total_runtime_s":239192,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.947368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-08835bcadb89a78cfc2b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-781978574287f8e6a11065c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":965,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[482];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":965,"metadata":{"avgScore":0.4433841810461951,"avg_score":0.4433841810461951,"avg_score_percent":44.33841810461951,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.07,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":585468127,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2068701,"passRate":0.3,"pass_rate":0.3,"pass_rate_percent":30.0,"passes":15,"runs":50,"runtime":47507,"source_flags":[],"source_harness":"codex","source_row_index":482,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":95.07,"totalDurationS":47507,"totalInputTokens":585468127,"totalOutputTokens":2068701,"total_cost_usd":95.07,"total_input_tokens":585468127,"total_output_tokens":2068701,"total_runtime_s":47507,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2b3f1703d683612f1ab6","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7818293b108b6a3f1ae60ec8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.17105263157894737,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":545,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[272];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":545,"metadata":{"avgScore":0.3850041500197,"avg_score":0.3850041500197,"avg_score_percent":38.50041500197,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":57,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":287164350,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1082713,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":151697,"source_flags":[],"source_harness":"codex","source_row_index":272,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":57,"totalDurationS":151697,"totalInputTokens":287164350,"totalOutputTokens":1082713,"total_cost_usd":57,"total_input_tokens":287164350,"total_output_tokens":1082713,"total_runtime_s":151697,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.105263157894736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ace108799c626aecd83c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-77b454699f4866ffe99e2df2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3262411347517731,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":855,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[427];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":855,"metadata":{"avgScore":0.5206445104262789,"avg_score":0.5206445104262789,"avg_score_percent":52.064451042627894,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":212.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1305756331,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":3877107,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":46,"runs":141,"runtime":189252,"source_flags":[],"source_harness":"codex","source_row_index":427,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":212.65,"totalDurationS":189252,"totalInputTokens":1305756331,"totalOutputTokens":3877107,"total_cost_usd":212.65,"total_input_tokens":1305756331,"total_output_tokens":3877107,"total_runtime_s":189252,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.62411347517731,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4d2260e47f398016236a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-76b7cc72c5f3b3a7ddcfd699","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29850746268656714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1251,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[625];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1251,"metadata":{"avgScore":0.6092947556539312,"avg_score":0.6092947556539312,"avg_score_percent":60.92947556539312,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":19.45,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":90618587,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":452393,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":67,"runtime":67571,"source_flags":[],"source_harness":"codex","source_row_index":625,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":19.45,"totalDurationS":67571,"totalInputTokens":90618587,"totalOutputTokens":452393,"total_cost_usd":19.45,"total_input_tokens":90618587,"total_output_tokens":452393,"total_runtime_s":67571,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.850746268656714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3f70285384c840d7270e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-75704703d97d1e20d637f22f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1095,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[547];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1095,"metadata":{"avgScore":0.1505198327191624,"avg_score":0.1505198327191624,"avg_score_percent":15.051983271916239,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":189.81,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1321242146,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2441114,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":30,"runtime":79440,"source_flags":[],"source_harness":"codex","source_row_index":547,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":189.81,"totalDurationS":79440,"totalInputTokens":1321242146,"totalOutputTokens":2441114,"total_cost_usd":189.81,"total_input_tokens":1321242146,"total_output_tokens":2441114,"total_runtime_s":79440,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a96d2d3712a89f3eee24","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6a13bfa6c82952a17bb96cff","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4925373134328358,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1197,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[598];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1197,"metadata":{"avgScore":0.7562644213793285,"avg_score":0.7562644213793285,"avg_score_percent":75.62644213793284,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.71,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":359705378,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1633460,"passRate":0.4925373134328358,"pass_rate":0.4925373134328358,"pass_rate_percent":49.25373134328358,"passes":33,"runs":67,"runtime":69711,"source_flags":[],"source_harness":"codex","source_row_index":598,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":68.71,"totalDurationS":69711,"totalInputTokens":359705378,"totalOutputTokens":1633460,"total_cost_usd":68.71,"total_input_tokens":359705378,"total_output_tokens":1633460,"total_runtime_s":69711,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.25373134328358,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f9673e3cd83ad7729642","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6639dd12917a210eec8dc8d5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":669,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[334];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":669,"metadata":{"avgScore":0.08178194518550498,"avg_score":0.08178194518550498,"avg_score_percent":8.178194518550498,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4.31,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":17611367,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":48073,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":10610,"source_flags":[],"source_harness":"codex","source_row_index":334,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":4.31,"totalDurationS":10610,"totalInputTokens":17611367,"totalOutputTokens":48073,"total_cost_usd":4.31,"total_input_tokens":17611367,"total_output_tokens":48073,"total_runtime_s":10610,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d6ebdefb3ce5b89314d5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-65d87e1330caf89dffc9e1a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18439716312056736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":911,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[455];split=unlicensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":911,"metadata":{"avgScore":0.40056329147969094,"avg_score":0.40056329147969094,"avg_score_percent":40.056329147969095,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.72,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":223199374,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":996976,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":140221,"source_flags":[],"source_harness":"codex","source_row_index":455,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":44.72,"totalDurationS":140221,"totalInputTokens":223199374,"totalOutputTokens":996976,"total_cost_usd":44.72,"total_input_tokens":223199374,"total_output_tokens":996976,"total_runtime_s":140221,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.439716312056735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cf237a7adf92fde4ccbb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5b415e98485b0d1afe527329","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18439716312056736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1383,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[691];split=unlicensed/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1383,"metadata":{"avgScore":0.40056329147969094,"avg_score":0.40056329147969094,"avg_score_percent":40.056329147969095,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.72,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":223199374,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":996976,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":140221,"source_flags":[],"source_harness":"codex","source_row_index":691,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":44.72,"totalDurationS":140221,"totalInputTokens":223199374,"totalOutputTokens":996976,"total_cost_usd":44.72,"total_input_tokens":223199374,"total_output_tokens":996976,"total_runtime_s":140221,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.439716312056735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e40a3d8fdf37dc8c2093","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5878e5fd66939f93605c0e11","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2727272727272727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":147,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[73];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":147,"metadata":{"avgScore":0.41712583069964326,"avg_score":0.41712583069964326,"avg_score_percent":41.712583069964325,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":351830472,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1354167,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":15,"runs":55,"runtime":81285,"source_flags":[],"source_harness":"codex","source_row_index":73,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":61.13,"totalDurationS":81285,"totalInputTokens":351830472,"totalOutputTokens":1354167,"total_cost_usd":61.13,"total_input_tokens":351830472,"total_output_tokens":1354167,"total_runtime_s":81285,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-693d2f3986db2a437107","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-55bed99f5d7aac3dc4fb4cc3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1137,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[568];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1137,"metadata":{"avgScore":0.09622840053229612,"avg_score":0.09622840053229612,"avg_score_percent":9.622840053229611,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":5.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":29419495,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":127636,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":14261,"source_flags":[],"source_harness":"codex","source_row_index":568,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":5.65,"totalDurationS":14261,"totalInputTokens":29419495,"totalOutputTokens":127636,"total_cost_usd":5.65,"total_input_tokens":29419495,"total_output_tokens":127636,"total_runtime_s":14261,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-869dcb23985de9267760","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5564be1af8e0056952b83d44","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.17105263157894737,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":73,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[36];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":73,"metadata":{"avgScore":0.3850041500197,"avg_score":0.3850041500197,"avg_score_percent":38.50041500197,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":57,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":287164350,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1082713,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":151697,"source_flags":[],"source_harness":"codex","source_row_index":36,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":57,"totalDurationS":151697,"totalInputTokens":287164350,"totalOutputTokens":1082713,"total_cost_usd":57,"total_input_tokens":287164350,"total_output_tokens":1082713,"total_runtime_s":151697,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.105263157894736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f056655e02b1b1f2e1fc","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-54d0f88e53ec1d3c018b9a29","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.43283582089552236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1201,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[600];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1201,"metadata":{"avgScore":0.743547028673177,"avg_score":0.743547028673177,"avg_score_percent":74.3547028673177,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":541671893,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":2191134,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":67,"runtime":78518,"source_flags":[],"source_harness":"codex","source_row_index":600,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.67,"totalDurationS":78518,"totalInputTokens":541671893,"totalOutputTokens":2191134,"total_cost_usd":95.67,"total_input_tokens":541671893,"total_output_tokens":2191134,"total_runtime_s":78518,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.28358208955223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2187b508c82e43da10f2","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-46dedfe2be543e0bb80d27d3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":299,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[149];split=full/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":299,"metadata":{"avgScore":0.10395747737317527,"avg_score":0.10395747737317527,"avg_score_percent":10.395747737317528,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":23.91,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":132807946,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":319379,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":56504,"source_flags":[],"source_harness":"codex","source_row_index":149,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":23.91,"totalDurationS":56504,"totalInputTokens":132807946,"totalOutputTokens":319379,"total_cost_usd":23.91,"total_input_tokens":132807946,"total_output_tokens":319379,"total_runtime_s":56504,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ffc9621100d90940a445","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4696e27490117893a78b959a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22388059701492538,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":445,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[222];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":445,"metadata":{"avgScore":0.48273103955698327,"avg_score":0.48273103955698327,"avg_score_percent":48.273103955698325,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":10.56,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":46726983,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":263717,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":67,"runtime":50813,"source_flags":[],"source_harness":"codex","source_row_index":222,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":10.56,"totalDurationS":50813,"totalInputTokens":46726983,"totalOutputTokens":263717,"total_cost_usd":10.56,"total_input_tokens":46726983,"total_output_tokens":263717,"total_runtime_s":50813,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.388059701492537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3a78ad47e8989f8707b5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-41758eee66be105ffea44c0a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.06,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1049,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[524];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1049,"metadata":{"avgScore":0.24564591975351033,"avg_score":0.24564591975351033,"avg_score_percent":24.564591975351032,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":6.8,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":28407280,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":188827,"passRate":0.06,"pass_rate":0.06,"pass_rate_percent":6.0,"passes":3,"runs":50,"runtime":15568,"source_flags":[],"source_harness":"codex","source_row_index":524,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":6.8,"totalDurationS":15568,"totalInputTokens":28407280,"totalOutputTokens":188827,"total_cost_usd":6.8,"total_input_tokens":28407280,"total_output_tokens":188827,"total_runtime_s":15568,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8935e6e29d11ed4efd60","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-39b089a37a484bd595221d03","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3026315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":493,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[246];split=full/overall;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":493,"metadata":{"avgScore":0.49414814826021464,"avg_score":0.49414814826021464,"avg_score_percent":49.41481482602146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235.12,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1435261861,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":4162374,"passRate":0.3026315789473684,"pass_rate":0.3026315789473684,"pass_rate_percent":30.263157894736842,"passes":46,"runs":152,"runtime":238042,"source_flags":[],"source_harness":"codex","source_row_index":246,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":235.12,"totalDurationS":238042,"totalInputTokens":1435261861,"totalOutputTokens":4162374,"total_cost_usd":235.12,"total_input_tokens":1435261861,"total_output_tokens":4162374,"total_runtime_s":238042,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.263157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8240007be7b7586d66da","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-25f804276ca8334d31523ead","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":655,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[327];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":655,"metadata":{"avgScore":0.1268451543325831,"avg_score":0.1268451543325831,"avg_score_percent":12.684515433258309,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":19.08,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":100638252,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":306689,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":8,"runtime":39399,"source_flags":[],"source_harness":"codex","source_row_index":327,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":8,"tokens":null,"totalCostUsd":19.08,"totalDurationS":39399,"totalInputTokens":100638252,"totalOutputTokens":306689,"total_cost_usd":19.08,"total_input_tokens":100638252,"total_output_tokens":306689,"total_runtime_s":39399,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2161cf61612627c651f1","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1fe94478d5eff159732816ec","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":647,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[323];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":647,"metadata":{"avgScore":0.15451296049520938,"avg_score":0.15451296049520938,"avg_score_percent":15.451296049520938,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.47,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":129505530,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":285267,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":48790,"source_flags":[],"source_harness":"codex","source_row_index":323,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":22.47,"totalDurationS":48790,"totalInputTokens":129505530,"totalOutputTokens":285267,"total_cost_usd":22.47,"total_input_tokens":129505530,"total_output_tokens":285267,"total_runtime_s":48790,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9b4815ce70a7e05df641","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1ca4f98b812e597d1df1ef52","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3582089552238806,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1223,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[611];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1223,"metadata":{"avgScore":0.6736742823508564,"avg_score":0.6736742823508564,"avg_score_percent":67.36742823508564,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":228811254,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1065059,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":67,"runtime":60584,"source_flags":[],"source_harness":"codex","source_row_index":611,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":46.18,"totalDurationS":60584,"totalInputTokens":228811254,"totalOutputTokens":1065059,"total_cost_usd":46.18,"total_input_tokens":228811254,"total_output_tokens":1065059,"total_runtime_s":60584,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.82089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7584e068193f272d558d","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1a3df7abcb8e6ccadb683560","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.12727272727272726,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":195,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[97];split=full/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":195,"metadata":{"avgScore":0.3007389590993101,"avg_score":0.3007389590993101,"avg_score_percent":30.073895909931007,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":18.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":85588559,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":355569,"passRate":0.12727272727272726,"pass_rate":0.12727272727272726,"pass_rate_percent":12.727272727272727,"passes":7,"runs":55,"runtime":32130,"source_flags":[],"source_harness":"codex","source_row_index":97,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":18.13,"totalDurationS":32130,"totalInputTokens":85588559,"totalOutputTokens":355569,"total_cost_usd":18.13,"total_input_tokens":85588559,"total_output_tokens":355569,"total_runtime_s":32130,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.727272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f223e022d3a43cfdbc28","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0eb2c8521d7b59bf6f7ee743","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":609,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[304];split=licensed;harness=codex;model=GPT-5.6-Luna;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":609,"metadata":{"avgScore":0.2343520785550317,"avg_score":0.2343520785550317,"avg_score_percent":23.435207855503172,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":11.06,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":56109934,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":143436,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":22844,"source_flags":[],"source_harness":"codex","source_row_index":304,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":11.06,"totalDurationS":22844,"totalInputTokens":56109934,"totalOutputTokens":143436,"total_cost_usd":11.06,"total_input_tokens":56109934,"total_output_tokens":143436,"total_runtime_s":22844,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-129baefa84f973850c40","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0d55cbe815c62c6801933073","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1119,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[559];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1119,"metadata":{"avgScore":0.11752938705842898,"avg_score":0.11752938705842898,"avg_score_percent":11.7529387058429,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":94.89,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":665277010,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1151373,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":71143,"source_flags":[],"source_harness":"codex","source_row_index":559,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":94.89,"totalDurationS":71143,"totalInputTokens":665277010,"totalOutputTokens":1151373,"total_cost_usd":94.89,"total_input_tokens":665277010,"total_output_tokens":1151373,"total_runtime_s":71143,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d4963f5968e085f22bc6","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0c104dc565e5cfb8dd74268a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2894736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":19,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[9];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":19,"metadata":{"avgScore":0.5037352082410917,"avg_score":0.5037352082410917,"avg_score_percent":50.37352082410918,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":390.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":2499272937,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":6752715,"passRate":0.2894736842105263,"pass_rate":0.2894736842105263,"pass_rate_percent":28.947368421052634,"passes":44,"runs":149,"runtime":239192,"source_flags":[],"source_harness":"codex","source_row_index":9,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":390.61,"totalDurationS":239192,"totalInputTokens":2499272937,"totalOutputTokens":6752715,"total_cost_usd":390.61,"total_input_tokens":2499272937,"total_output_tokens":6752715,"total_runtime_s":239192,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.947368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e606570e80e572aab6b0","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-03a2db6248b975db950d9292","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.11428571428571428,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":803,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[401];split=linux_only;harness=codex;model=GPT-5.6-Luna;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":803,"metadata":{"avgScore":0.3361045019913022,"avg_score":0.3361045019913022,"avg_score_percent":33.61045019913022,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.85,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":56323564,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":407087,"passRate":0.11428571428571428,"pass_rate":0.11428571428571428,"pass_rate_percent":11.428571428571429,"passes":12,"runs":105,"runtime":33091,"source_flags":[],"source_harness":"codex","source_row_index":401,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":12.85,"totalDurationS":33091,"totalInputTokens":56323564,"totalOutputTokens":407087,"total_cost_usd":12.85,"total_input_tokens":56323564,"total_output_tokens":407087,"total_runtime_s":33091,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.428571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8a66b06d39ad0c46fa75","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-009ce552f5950e7d62d3d6b2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3026315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":21,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[10];split=full;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":21,"metadata":{"avgScore":0.49414814826021464,"avg_score":0.49414814826021464,"avg_score_percent":49.41481482602146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235.12,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":1435261861,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":4162374,"passRate":0.3026315789473684,"pass_rate":0.3026315789473684,"pass_rate_percent":30.263157894736842,"passes":46,"runs":152,"runtime":238042,"source_flags":[],"source_harness":"codex","source_row_index":10,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":235.12,"totalDurationS":238042,"totalInputTokens":1435261861,"totalOutputTokens":4162374,"total_cost_usd":235.12,"total_input_tokens":1435261861,"total_output_tokens":4162374,"total_runtime_s":238042,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.263157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d2e239f4fe3627d241cb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0095275f439366ceca18a45f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":981,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[490];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Luna;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":981,"metadata":{"avgScore":0.4177055625187485,"avg_score":0.4177055625187485,"avg_score_percent":41.77055625187485,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":53.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":299155581,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":1250311,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":14,"runs":50,"runtime":51596,"source_flags":[],"source_harness":"codex","source_row_index":490,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":53.18,"totalDurationS":51596,"totalInputTokens":299155581,"totalOutputTokens":1250311,"total_cost_usd":53.18,"total_input_tokens":299155581,"total_output_tokens":1250311,"total_runtime_s":51596,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-96a6faa21ef0190f6247","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0078961745ba978a118d4141","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Luna","modelRef":"GPT-5.6-Luna","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29850746268656714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27917621169de2d353b8001427011c50c6b3334e","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":415,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[207];split=full/near-term;harness=codex;model=GPT-5.6-Luna;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":415,"metadata":{"avgScore":0.6092947556539312,"avg_score":0.6092947556539312,"avg_score_percent":60.92947556539312,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":19.45,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":90618587,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Luna","output_tokens":452393,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":67,"runtime":67571,"source_flags":[],"source_harness":"codex","source_row_index":207,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":19.45,"totalDurationS":67571,"totalInputTokens":90618587,"totalOutputTokens":452393,"total_cost_usd":19.45,"total_input_tokens":90618587,"total_output_tokens":452393,"total_runtime_s":67571,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.850746268656714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-00ba7e89d1db699db34f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fded58569e7e478e7cd940c9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1261806657735013,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":276,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[137];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":276,"metadata":{"avgScore":0.1261806657735013,"avg_score":0.1261806657735013,"avg_score_percent":12.61806657735013,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":96.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":111296708,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":408600.3333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":74,"runtime":109927.33333333333,"source_flags":[],"source_harness":"codex","source_row_index":137,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":96.13,"totalDurationS":109927.33333333333,"totalInputTokens":111296708,"totalOutputTokens":408600.3333333333,"total_cost_usd":96.13,"total_input_tokens":111296708,"total_output_tokens":408600.3333333333,"total_runtime_s":109927.33333333333,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.61806657735013,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92441f6744dbee2d628e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fadea1877e295ac21ba9cfeb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18922557053100295,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1088,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[543];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1088,"metadata":{"avgScore":0.18922557053100295,"avg_score":0.18922557053100295,"avg_score_percent":18.922557053100295,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":183.03,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":231236782,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":657382,"passRate":0.05,"pass_rate":0.05,"pass_rate_percent":5.0,"passes":2,"runs":59,"runtime":84648,"source_flags":[],"source_harness":"codex","source_row_index":543,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":183.03,"totalDurationS":84648,"totalInputTokens":231236782,"totalOutputTokens":657382,"total_cost_usd":183.03,"total_input_tokens":231236782,"total_output_tokens":657382,"total_runtime_s":84648,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.922557053100295,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eccae3ef17022b305311","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fac7ddb6d5db33448ed711cc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5363939737297058,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":710,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[354];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":710,"metadata":{"avgScore":0.5363939737297058,"avg_score":0.5363939737297058,"avg_score_percent":53.63939737297058,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446.38,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":443472815.3333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2515905.5,"passRate":0.2857142857142857,"pass_rate":0.2857142857142857,"pass_rate_percent":28.57142857142857,"passes":60,"runs":209,"runtime":179145.1965,"source_flags":[],"source_harness":"codex","source_row_index":354,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":446.38,"totalDurationS":179145.1965,"totalInputTokens":443472815.3333333,"totalOutputTokens":2515905.5,"total_cost_usd":446.38,"total_input_tokens":443472815.3333333,"total_output_tokens":2515905.5,"total_runtime_s":179145.1965,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.63939737297058,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e753b32e573829d86808","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fa7d64ef7743bbe96513439a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7829241382323966,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":358,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[178];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":358,"metadata":{"avgScore":0.7829241382323966,"avg_score":0.7829241382323966,"avg_score_percent":78.29241382323966,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":210.38,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":189978516.33333334,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1282200,"passRate":0.4552238805970149,"pass_rate":0.4552238805970149,"pass_rate_percent":45.52238805970149,"passes":61,"runs":134,"runtime":125006.5,"source_flags":[],"source_harness":"codex","source_row_index":178,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":210.38,"totalDurationS":125006.5,"totalInputTokens":189978516.33333334,"totalOutputTokens":1282200,"total_cost_usd":210.38,"total_input_tokens":189978516.33333334,"total_output_tokens":1282200,"total_runtime_s":125006.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":78.29241382323966,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dcc36c96d5f0aab08d7f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f85f98c24b26a6fdc5418b43","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.531374869029037,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":712,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[355];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":712,"metadata":{"avgScore":0.531374869029037,"avg_score":0.531374869029037,"avg_score_percent":53.137486902903696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":279.31,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":300085777.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1422241,"passRate":0.2857142857142857,"pass_rate":0.2857142857142857,"pass_rate_percent":28.57142857142857,"passes":57,"runs":207,"runtime":171176.44199999998,"source_flags":[],"source_harness":"codex","source_row_index":355,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":279.31,"totalDurationS":171176.44199999998,"totalInputTokens":300085777.5,"totalOutputTokens":1422241,"total_cost_usd":279.31,"total_input_tokens":300085777.5,"total_output_tokens":1422241,"total_runtime_s":171176.44199999998,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.137486902903696,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0d1b4affaefba654e3cf","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ef1aca12a05d260dfb61c72b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.775482331051408,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1196,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[597];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1196,"metadata":{"avgScore":0.775482331051408,"avg_score":0.775482331051408,"avg_score_percent":77.5482331051408,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":172.04,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155266763,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1008998,"passRate":0.4701492537313433,"pass_rate":0.4701492537313433,"pass_rate_percent":47.01492537313433,"passes":62,"runs":132,"runtime":120620.5,"source_flags":[],"source_harness":"codex","source_row_index":597,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":172.04,"totalDurationS":120620.5,"totalInputTokens":155266763,"totalOutputTokens":1008998,"total_cost_usd":172.04,"total_input_tokens":155266763,"total_output_tokens":1008998,"total_runtime_s":120620.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":77.5482331051408,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2e9fb829d72a7a2b7195","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e94bd87d0a79e5b66d6fe09a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5595483966613978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":838,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[418];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":838,"metadata":{"avgScore":0.5595483966613978,"avg_score":0.5595483966613978,"avg_score_percent":55.954839666139776,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":656.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":652364022.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":3478380.3333333335,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":92,"runs":280,"runtime":291953.19649999996,"source_flags":[],"source_harness":"codex","source_row_index":418,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":656.72,"totalDurationS":291953.19649999996,"totalInputTokens":652364022.1666666,"totalOutputTokens":3478380.3333333335,"total_cost_usd":656.72,"total_input_tokens":652364022.1666666,"total_output_tokens":3478380.3333333335,"total_runtime_s":291953.19649999996,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":55.954839666139776,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-31f441dc7594eb3e659e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d90f6237b6e09c0b7ec130fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4681048909575152,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":876,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[437];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":876,"metadata":{"avgScore":0.4681048909575152,"avg_score":0.4681048909575152,"avg_score_percent":46.81048909575152,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":223.43,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":230074698.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1175061.3333333333,"passRate":0.2541371158392435,"pass_rate":0.2541371158392435,"pass_rate_percent":25.413711583924346,"passes":72,"runs":278,"runtime":242446.52566666668,"source_flags":[],"source_harness":"codex","source_row_index":437,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":223.43,"totalDurationS":242446.52566666668,"totalInputTokens":230074698.16666666,"totalOutputTokens":1175061.3333333333,"total_cost_usd":223.43,"total_input_tokens":230074698.16666666,"total_output_tokens":1175061.3333333333,"total_runtime_s":242446.52566666668,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.81048909575152,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a4ff8911da2279dbf5cb","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d442fd1e64efb1bb744ec407","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5210026942935828,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":10,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[4];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":10,"metadata":{"avgScore":0.5210026942935828,"avg_score":0.5210026942935828,"avg_score_percent":52.10026942935828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":518.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":539877934.6666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2196349.5,"passRate":0.29495614035087714,"pass_rate":0.29495614035087714,"pass_rate_percent":29.495614035087712,"passes":87,"runs":302,"runtime":328693.94200000004,"source_flags":[],"source_harness":"codex","source_row_index":4,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":518.72,"totalDurationS":328693.94200000004,"totalInputTokens":539877934.6666666,"totalOutputTokens":2196349.5,"total_cost_usd":518.72,"total_input_tokens":539877934.6666666,"total_output_tokens":2196349.5,"total_runtime_s":328693.94200000004,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.10026942935828,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92c83805b680bd992db7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cd796305a7e11eae03bfa53c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5491172987872457,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1314,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[656];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1314,"metadata":{"avgScore":0.5491172987872457,"avg_score":0.5491172987872457,"avg_score_percent":54.91172987872457,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":525.89,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":543170921.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2693839,"passRate":0.32978723404255317,"pass_rate":0.32978723404255317,"pass_rate_percent":32.97872340425532,"passes":91,"runs":278,"runtime":282712.47,"source_flags":[],"source_harness":"codex","source_row_index":656,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":525.89,"totalDurationS":282712.47,"totalInputTokens":543170921.5,"totalOutputTokens":2693839,"total_cost_usd":525.89,"total_input_tokens":543170921.5,"total_output_tokens":2693839,"total_runtime_s":282712.47,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.91172987872457,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1f61e0ecbadec7d8974e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cc0e2f39ccb2515302852530","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5430423343738866,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":846,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[422];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":846,"metadata":{"avgScore":0.5430423343738866,"avg_score":0.5430423343738866,"avg_score_percent":54.30423343738866,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":424.24,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":447187737.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2005900.5,"passRate":0.317966903073286,"pass_rate":0.317966903073286,"pass_rate_percent":31.7966903073286,"passes":87,"runs":280,"runtime":273534.44200000004,"source_flags":[],"source_harness":"codex","source_row_index":422,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":424.24,"totalDurationS":273534.44200000004,"totalInputTokens":447187737.1666667,"totalOutputTokens":2005900.5,"total_cost_usd":424.24,"total_input_tokens":447187737.1666667,"total_output_tokens":2005900.5,"total_runtime_s":273534.44200000004,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.30423343738866,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8bf972f95fcb349e269","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cbbbaba2d18cffb992ba8d55","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6677679137696316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":396,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[197];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":396,"metadata":{"avgScore":0.6677679137696316,"avg_score":0.6677679137696316,"avg_score_percent":66.77679137696316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.16,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":71044249.66666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":467625,"passRate":0.3930348258706468,"pass_rate":0.3930348258706468,"pass_rate_percent":39.30348258706468,"passes":53,"runs":134,"runtime":96179.83333333334,"source_flags":[],"source_harness":"codex","source_row_index":197,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":80.16,"totalDurationS":96179.83333333334,"totalInputTokens":71044249.66666667,"totalOutputTokens":467625,"total_cost_usd":80.16,"total_input_tokens":71044249.66666667,"total_output_tokens":467625,"total_runtime_s":96179.83333333334,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":66.77679137696316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c5cfd1374c4389ccd96","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c81b61bafcc67abc06836d3a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5430423343738866,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1318,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[658];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1318,"metadata":{"avgScore":0.5430423343738866,"avg_score":0.5430423343738866,"avg_score_percent":54.30423343738866,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":424.24,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":447187737.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2005900.5,"passRate":0.317966903073286,"pass_rate":0.317966903073286,"pass_rate_percent":31.7966903073286,"passes":87,"runs":280,"runtime":273534.44200000004,"source_flags":[],"source_harness":"codex","source_row_index":658,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":424.24,"totalDurationS":273534.44200000004,"totalInputTokens":447187737.1666667,"totalOutputTokens":2005900.5,"total_cost_usd":424.24,"total_input_tokens":447187737.1666667,"total_output_tokens":2005900.5,"total_runtime_s":273534.44200000004,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.30423343738866,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-412653b20d42c580ed34","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c7776423a3a9b2f319170d8f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.523728673169988,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":480,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[239];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":480,"metadata":{"avgScore":0.523728673169988,"avg_score":0.523728673169988,"avg_score_percent":52.37286731699879,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":575.68,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":594028867,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2890301.5,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":91,"runs":299,"runtime":340491.97,"source_flags":[],"source_harness":"codex","source_row_index":239,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":575.68,"totalDurationS":340491.97,"totalInputTokens":594028867,"totalOutputTokens":2890301.5,"total_cost_usd":575.68,"total_input_tokens":594028867,"total_output_tokens":2890301.5,"total_runtime_s":340491.97,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.37286731699879,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7b88877a757df4721cd3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c3761fefb0ef48e0b5f90583","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5276278004520454,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":476,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[237];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":476,"metadata":{"avgScore":0.5276278004520454,"avg_score":0.5276278004520454,"avg_score_percent":52.762780045204536,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1084.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1095236840.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":5443524.666666667,"passRate":0.29605263157894735,"pass_rate":0.29605263157894735,"pass_rate_percent":29.605263157894733,"passes":89,"runs":307,"runtime":373587.878,"source_flags":[],"source_harness":"codex","source_row_index":237,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":1084.09,"totalDurationS":373587.878,"totalInputTokens":1095236840.8333333,"totalOutputTokens":5443524.666666667,"total_cost_usd":1084.09,"total_input_tokens":1095236840.8333333,"total_output_tokens":5443524.666666667,"total_runtime_s":373587.878,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.762780045204536,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4be72248a4dd1717b506","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c22ceb4c6b663d1e767d2cd0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5361628463309976,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":2,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[0];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":2,"metadata":{"avgScore":0.5361628463309976,"avg_score":0.5361628463309976,"avg_score_percent":53.61628463309977,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":771.62,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":762936773.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":3798417.8333333335,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":93,"runs":301,"runtime":340729.69649999996,"source_flags":[],"source_harness":"codex","source_row_index":0,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":771.62,"totalDurationS":340729.69649999996,"totalInputTokens":762936773.1666666,"totalOutputTokens":3798417.8333333335,"total_cost_usd":771.62,"total_input_tokens":762936773.1666666,"total_output_tokens":3798417.8333333335,"total_runtime_s":340729.69649999996,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.61628463309977,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd76c990ae79f83b1e9f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bda5768f50630b62e57a516a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4451610959401981,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":134,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[66];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":134,"metadata":{"avgScore":0.4451610959401981,"avg_score":0.4451610959401981,"avg_score_percent":44.51610959401981,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":157.67,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":142119530.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":732332.5,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":29,"runs":109,"runtime":107957.442,"source_flags":[],"source_harness":"codex","source_row_index":66,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":157.67,"totalDurationS":107957.442,"totalInputTokens":142119530.5,"totalOutputTokens":732332.5,"total_cost_usd":157.67,"total_input_tokens":142119530.5,"total_output_tokens":732332.5,"total_runtime_s":107957.442,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.51610959401981,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aa3ea65a65f578d678f0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bb8ee68f24773ba0950c8432","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.433638270435078,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":974,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[486];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":974,"metadata":{"avgScore":0.433638270435078,"avg_score":0.433638270435078,"avg_score_percent":43.363827043507804,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":138.39,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":124762948,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":846635.5,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":28,"runs":99,"runtime":81367.47,"source_flags":[],"source_harness":"codex","source_row_index":486,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":138.39,"totalDurationS":81367.47,"totalInputTokens":124762948,"totalOutputTokens":846635.5,"total_cost_usd":138.39,"total_input_tokens":124762948,"total_output_tokens":846635.5,"total_runtime_s":81367.47,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.363827043507804,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-29ede4ad598a2ac05138","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b734c120e9046d4447ea66a9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44076492497958797,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":968,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[483];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":968,"metadata":{"avgScore":0.44076492497958797,"avg_score":0.44076492497958797,"avg_score_percent":44.076492497958796,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":239.3,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":217198822,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1560829.1666666667,"passRate":0.26,"pass_rate":0.26,"pass_rate_percent":26.0,"passes":25,"runs":100,"runtime":74690.378,"source_flags":[],"source_harness":"codex","source_row_index":483,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":239.3,"totalDurationS":74690.378,"totalInputTokens":217198822,"totalOutputTokens":1560829.1666666667,"total_cost_usd":239.3,"total_input_tokens":217198822,"total_output_tokens":1560829.1666666667,"total_runtime_s":74690.378,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.076492497958796,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-200b8cffe76c76ac0db1","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b35d633e987d2bafa3c175c6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5276278004520454,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":4,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[1];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":4,"metadata":{"avgScore":0.5276278004520454,"avg_score":0.5276278004520454,"avg_score_percent":52.762780045204536,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1084.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1095236840.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":5443524.666666667,"passRate":0.29605263157894735,"pass_rate":0.29605263157894735,"pass_rate_percent":29.605263157894733,"passes":89,"runs":307,"runtime":373587.878,"source_flags":[],"source_harness":"codex","source_row_index":1,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":1084.09,"totalDurationS":373587.878,"totalInputTokens":1095236840.8333333,"totalOutputTokens":5443524.666666667,"total_cost_usd":1084.09,"total_input_tokens":1095236840.8333333,"total_output_tokens":5443524.666666667,"total_runtime_s":373587.878,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.762780045204536,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4938ed64fd76fdada8e8","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b355b107d2b4368cde792f56","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7882473228071811,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":356,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[177];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":356,"metadata":{"avgScore":0.7882473228071811,"avg_score":0.7882473228071811,"avg_score_percent":78.82473228071811,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":322.09,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":304163441.3333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1816516,"passRate":0.47761194029850745,"pass_rate":0.47761194029850745,"pass_rate_percent":47.76119402985074,"passes":63,"runs":134,"runtime":145737.33333333334,"source_flags":[],"source_harness":"codex","source_row_index":177,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":322.09,"totalDurationS":145737.33333333334,"totalInputTokens":304163441.3333333,"totalOutputTokens":1816516,"total_cost_usd":322.09,"total_input_tokens":304163441.3333333,"total_output_tokens":1816516,"total_runtime_s":145737.33333333334,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":78.82473228071811,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-de17a809f42dc533532e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b323823a7077827d730553d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6677679137696316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1232,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[615];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1232,"metadata":{"avgScore":0.6677679137696316,"avg_score":0.6677679137696316,"avg_score_percent":66.77679137696316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.16,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":71044249.66666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":467625,"passRate":0.3930348258706468,"pass_rate":0.3930348258706468,"pass_rate_percent":39.30348258706468,"passes":53,"runs":134,"runtime":96179.83333333334,"source_flags":[],"source_harness":"codex","source_row_index":615,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":80.16,"totalDurationS":96179.83333333334,"totalInputTokens":71044249.66666667,"totalOutputTokens":467625,"total_cost_usd":80.16,"total_input_tokens":71044249.66666667,"total_output_tokens":467625,"total_runtime_s":96179.83333333334,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":66.77679137696316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bdd9db644125a4ce3589","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-af6c8eb5f50556c16dddf401","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7561914737576828,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1200,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[599];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1200,"metadata":{"avgScore":0.7561914737576828,"avg_score":0.7561914737576828,"avg_score_percent":75.61914737576828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":136.75,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":124228619.66666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":787503.5,"passRate":0.45273631840796025,"pass_rate":0.45273631840796025,"pass_rate_percent":45.27363184079603,"passes":60,"runs":134,"runtime":123361.5,"source_flags":[],"source_harness":"codex","source_row_index":599,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":136.75,"totalDurationS":123361.5,"totalInputTokens":124228619.66666667,"totalOutputTokens":787503.5,"total_cost_usd":136.75,"total_input_tokens":124228619.66666667,"total_output_tokens":787503.5,"total_runtime_s":123361.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":75.61914737576828,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-29b4db3bc036d800f890","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ae9c99a6259707a2844d70d0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.12315914835731925,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1116,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[557];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1116,"metadata":{"avgScore":0.12315914835731925,"avg_score":0.12315914835731925,"avg_score_percent":12.315914835731926,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":85.84,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":102937640,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":356578.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":58,"runtime":87469.83333333333,"source_flags":[],"source_harness":"codex","source_row_index":557,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":85.84,"totalDurationS":87469.83333333333,"totalInputTokens":102937640,"totalOutputTokens":356578.8333333333,"total_cost_usd":85.84,"total_input_tokens":102937640,"total_output_tokens":356578.8333333333,"total_runtime_s":87469.83333333333,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.315914835731926,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e9391cf73a99e4d9ae2f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ae4d716a3ef4e3328980c416","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.512155695633506,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":720,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[359];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":720,"metadata":{"avgScore":0.512155695633506,"avg_score":0.512155695633506,"avg_score_percent":51.215569563350606,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":343.41,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":359092352,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1936844.5,"passRate":0.2714285714285714,"pass_rate":0.2714285714285714,"pass_rate_percent":27.142857142857142,"passes":55,"runs":206,"runtime":184402.47,"source_flags":[],"source_harness":"codex","source_row_index":359,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":343.41,"totalDurationS":184402.47,"totalInputTokens":359092352,"totalOutputTokens":1936844.5,"total_cost_usd":343.41,"total_input_tokens":359092352,"total_output_tokens":1936844.5,"total_runtime_s":184402.47,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.215569563350606,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e7b75cfda6d1329019e0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ac6522c275562d2fab85c165","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4301432864776319,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":140,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[69];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":140,"metadata":{"avgScore":0.4301432864776319,"avg_score":0.4301432864776319,"avg_score_percent":43.014328647763186,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":164.83,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":153018341,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":942915.5,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":30,"runs":108,"runtime":107709.47,"source_flags":[],"source_harness":"codex","source_row_index":69,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":164.83,"totalDurationS":107709.47,"totalInputTokens":153018341,"totalOutputTokens":942915.5,"total_cost_usd":164.83,"total_input_tokens":153018341,"total_output_tokens":942915.5,"total_runtime_s":107709.47,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.014328647763186,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e4daf000a6f0156b0b05","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a3bd64d04637eef39ae007a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23849458053696015,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":606,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[302];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":606,"metadata":{"avgScore":0.23849458053696015,"avg_score":0.23849458053696015,"avg_score_percent":23.849458053696015,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":94.47,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":92690197.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":190449,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":55159.5,"source_flags":[],"source_harness":"codex","source_row_index":302,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":94.47,"totalDurationS":55159.5,"totalInputTokens":92690197.5,"totalOutputTokens":190449,"total_cost_usd":94.47,"total_input_tokens":92690197.5,"total_output_tokens":190449,"total_runtime_s":55159.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.849458053696015,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a56230b424cdcd89ee23","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-9e13cf05da4f2a586bfc3b7a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1644048385293532,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":256,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[127];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":256,"metadata":{"avgScore":0.1644048385293532,"avg_score":0.1644048385293532,"avg_score_percent":16.44048385293532,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":488.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":523793885.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2091747,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":2,"runs":79,"runtime":154166.1666666667,"source_flags":[],"source_harness":"codex","source_row_index":127,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":488.13,"totalDurationS":154166.1666666667,"totalInputTokens":523793885.5,"totalOutputTokens":2091747,"total_cost_usd":488.13,"total_input_tokens":523793885.5,"total_output_tokens":2091747,"total_runtime_s":154166.1666666667,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.44048385293532,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd39ea6bd57cea97c314","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-97a1e8589d01e11ee1e2da1d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7882473228071811,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1192,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[595];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1192,"metadata":{"avgScore":0.7882473228071811,"avg_score":0.7882473228071811,"avg_score_percent":78.82473228071811,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":322.09,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":304163441.3333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1816516,"passRate":0.47761194029850745,"pass_rate":0.47761194029850745,"pass_rate_percent":47.76119402985074,"passes":63,"runs":134,"runtime":145737.33333333334,"source_flags":[],"source_harness":"codex","source_row_index":595,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":322.09,"totalDurationS":145737.33333333334,"totalInputTokens":304163441.3333333,"totalOutputTokens":1816516,"total_cost_usd":322.09,"total_input_tokens":304163441.3333333,"total_output_tokens":1816516,"total_runtime_s":145737.33333333334,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":78.82473228071811,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-36496853e17113c516df","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-94e233dd8465c75cff2eaa4d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23140343373334019,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":612,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[305];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":612,"metadata":{"avgScore":0.23140343373334019,"avg_score":0.23140343373334019,"avg_score_percent":23.14034337333402,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":24.69,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":21887578.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":93987.5,"passRate":0.045454545454545456,"pass_rate":0.045454545454545456,"pass_rate_percent":4.545454545454546,"passes":1,"runs":21,"runtime":28205.5,"source_flags":[],"source_harness":"codex","source_row_index":305,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":24.69,"totalDurationS":28205.5,"totalInputTokens":21887578.5,"totalOutputTokens":93987.5,"total_cost_usd":24.69,"total_input_tokens":21887578.5,"total_output_tokens":93987.5,"total_runtime_s":28205.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.14034337333402,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-395437a6688eceb910c9","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-9463536824de9af79701af65","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5595483966613978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1310,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[654];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1310,"metadata":{"avgScore":0.5595483966613978,"avg_score":0.5595483966613978,"avg_score_percent":55.954839666139776,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":656.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":652364022.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":3478380.3333333335,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":92,"runs":280,"runtime":291953.19649999996,"source_flags":[],"source_harness":"codex","source_row_index":654,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":656.72,"totalDurationS":291953.19649999996,"totalInputTokens":652364022.1666666,"totalOutputTokens":3478380.3333333335,"total_cost_usd":656.72,"total_input_tokens":652364022.1666666,"total_output_tokens":3478380.3333333335,"total_runtime_s":291953.19649999996,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":55.954839666139776,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-37f9510a36ca41bcbfb0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8d3e9d3d0215b8c91914cedb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19652298994758152,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1086,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[542];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1086,"metadata":{"avgScore":0.19652298994758152,"avg_score":0.19652298994758152,"avg_score_percent":19.65229899475815,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":227.28,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":273176548.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":948762,"passRate":0.06666666666666667,"pass_rate":0.06666666666666667,"pass_rate_percent":6.666666666666667,"passes":3,"runs":59,"runtime":89607,"source_flags":[],"source_harness":"codex","source_row_index":542,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":227.28,"totalDurationS":89607,"totalInputTokens":273176548.5,"totalOutputTokens":948762,"total_cost_usd":227.28,"total_input_tokens":273176548.5,"total_output_tokens":948762,"total_runtime_s":89607,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.65229899475815,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-079490741080ff65958f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8981d38a6333094148e0616e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23640261027768616,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":608,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[303];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":608,"metadata":{"avgScore":0.23640261027768616,"avg_score":0.23640261027768616,"avg_score_percent":23.640261027768616,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":114.9,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":110572751,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":320037.5,"passRate":0.045454545454545456,"pass_rate":0.045454545454545456,"pass_rate_percent":4.545454545454546,"passes":1,"runs":21,"runtime":48776.5,"source_flags":[],"source_harness":"codex","source_row_index":303,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":114.9,"totalDurationS":48776.5,"totalInputTokens":110572751,"totalOutputTokens":320037.5,"total_cost_usd":114.9,"total_input_tokens":110572751,"total_output_tokens":320037.5,"total_runtime_s":48776.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.640261027768616,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a8e2dfbd5f1776307cf6","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-812a879af8c83707e1805a6c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1868668857038979,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":252,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[125];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":252,"metadata":{"avgScore":0.1868668857038979,"avg_score":0.1868668857038979,"avg_score_percent":18.68668857038979,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":252.66,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":297522834,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1062195,"passRate":0.05263157894736842,"pass_rate":0.05263157894736842,"pass_rate_percent":5.263157894736842,"passes":3,"runs":75,"runtime":140665,"source_flags":[],"source_harness":"codex","source_row_index":125,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":252.66,"totalDurationS":140665,"totalInputTokens":297522834,"totalOutputTokens":1062195,"total_cost_usd":252.66,"total_input_tokens":297522834,"total_output_tokens":1062195,"total_runtime_s":140665,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.68668857038979,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9ce279cf82a0a230bd26","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7c91e22163e3fb1691d74292","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7561914737576828,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":364,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[181];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":364,"metadata":{"avgScore":0.7561914737576828,"avg_score":0.7561914737576828,"avg_score_percent":75.61914737576828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":136.75,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":124228619.66666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":787503.5,"passRate":0.45273631840796025,"pass_rate":0.45273631840796025,"pass_rate_percent":45.27363184079603,"passes":60,"runs":134,"runtime":123361.5,"source_flags":[],"source_harness":"codex","source_row_index":181,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":136.75,"totalDurationS":123361.5,"totalInputTokens":124228619.66666667,"totalOutputTokens":787503.5,"total_cost_usd":136.75,"total_input_tokens":124228619.66666667,"total_output_tokens":787503.5,"total_runtime_s":123361.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":75.61914737576828,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f7a40b05f46c04a811ba","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-78a64a0e1845259556f3f94b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7829241382323966,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1194,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[596];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1194,"metadata":{"avgScore":0.7829241382323966,"avg_score":0.7829241382323966,"avg_score_percent":78.29241382323966,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":210.38,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":189978516.33333334,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1282200,"passRate":0.4552238805970149,"pass_rate":0.4552238805970149,"pass_rate_percent":45.52238805970149,"passes":61,"runs":134,"runtime":125006.5,"source_flags":[],"source_harness":"codex","source_row_index":596,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":210.38,"totalDurationS":125006.5,"totalInputTokens":189978516.33333334,"totalOutputTokens":1282200,"total_cost_usd":210.38,"total_input_tokens":189978516.33333334,"total_output_tokens":1282200,"total_runtime_s":125006.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":78.29241382323966,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f21b3b4a13c20f429deb","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-774b8aeffef1a3b574546415","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.20737264567839306,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1080,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[539];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1080,"metadata":{"avgScore":0.20737264567839306,"avg_score":0.20737264567839306,"avg_score_percent":20.737264567839304,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":259.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":297768333.3333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1208400.8333333333,"passRate":0.06666666666666667,"pass_rate":0.06666666666666667,"pass_rate_percent":6.666666666666667,"passes":4,"runs":61,"runtime":97842,"source_flags":[],"source_harness":"codex","source_row_index":539,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":259.53,"totalDurationS":97842,"totalInputTokens":297768333.3333333,"totalOutputTokens":1208400.8333333333,"total_cost_usd":259.53,"total_input_tokens":297768333.3333333,"total_output_tokens":1208400.8333333333,"total_runtime_s":97842,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.737264567839304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a5c9628cddd66fe273c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-730ccdcab5ef217788be3d4b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5210026942935828,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":482,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[240];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":482,"metadata":{"avgScore":0.5210026942935828,"avg_score":0.5210026942935828,"avg_score_percent":52.10026942935828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":518.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":539877934.6666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2196349.5,"passRate":0.29495614035087714,"pass_rate":0.29495614035087714,"pass_rate_percent":29.495614035087712,"passes":87,"runs":302,"runtime":328693.94200000004,"source_flags":[],"source_harness":"codex","source_row_index":240,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":518.72,"totalDurationS":328693.94200000004,"totalInputTokens":539877934.6666666,"totalOutputTokens":2196349.5,"total_cost_usd":518.72,"total_input_tokens":539877934.6666666,"total_output_tokens":2196349.5,"total_runtime_s":328693.94200000004,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.10026942935828,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3402f0f61f687d608783","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-71c822a953d9555581893fa8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.39285408160616503,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":998,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[498];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":998,"metadata":{"avgScore":0.39285408160616503,"avg_score":0.39285408160616503,"avg_score_percent":39.285408160616505,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.9,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":59225023.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":389348.5,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":20,"runs":98,"runtime":65719.859,"source_flags":[],"source_harness":"codex","source_row_index":498,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":61.9,"totalDurationS":65719.859,"totalInputTokens":59225023.5,"totalOutputTokens":389348.5,"total_cost_usd":61.9,"total_input_tokens":59225023.5,"total_output_tokens":389348.5,"total_runtime_s":65719.859,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.285408160616505,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e867c1d48a0565eb4269","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6aa11bb93c32f07192589271","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4509751802373446,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":40,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[19];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":40,"metadata":{"avgScore":0.4509751802373446,"avg_score":0.4509751802373446,"avg_score_percent":45.09751802373446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":248.12,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":251962276.66666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1269048.8333333333,"passRate":0.2390350877192982,"pass_rate":0.2390350877192982,"pass_rate_percent":23.903508771929822,"passes":73,"runs":299,"runtime":270652.0256666667,"source_flags":[],"source_harness":"codex","source_row_index":19,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":248.12,"totalDurationS":270652.0256666667,"totalInputTokens":251962276.66666666,"totalOutputTokens":1269048.8333333333,"total_cost_usd":248.12,"total_input_tokens":251962276.66666666,"total_output_tokens":1269048.8333333333,"total_runtime_s":270652.0256666667,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.09751802373446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-725ad10fcf74a6907dbb","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-64435483a543f951c96fd396","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.523728673169988,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":8,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[3];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":8,"metadata":{"avgScore":0.523728673169988,"avg_score":0.523728673169988,"avg_score_percent":52.37286731699879,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":575.68,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":594028867,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2890301.5,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":91,"runs":299,"runtime":340491.97,"source_flags":[],"source_harness":"codex","source_row_index":3,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":575.68,"totalDurationS":340491.97,"totalInputTokens":594028867,"totalOutputTokens":2890301.5,"total_cost_usd":575.68,"total_input_tokens":594028867,"total_output_tokens":2890301.5,"total_runtime_s":340491.97,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.37286731699879,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9fac2f7d1d389e3099bb","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-60a46a4b17e13217cdeb0036","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5473632833709104,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":844,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[421];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":844,"metadata":{"avgScore":0.5473632833709104,"avg_score":0.5473632833709104,"avg_score_percent":54.736328337091045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":903.28,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":921100178.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":4976045.666666667,"passRate":0.3191489361702128,"pass_rate":0.3191489361702128,"pass_rate_percent":31.914893617021278,"passes":89,"runs":284,"runtime":311864.54466666665,"source_flags":[],"source_harness":"codex","source_row_index":421,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":903.28,"totalDurationS":311864.54466666665,"totalInputTokens":921100178.5,"totalOutputTokens":4976045.666666667,"total_cost_usd":903.28,"total_input_tokens":921100178.5,"total_output_tokens":4976045.666666667,"total_runtime_s":311864.54466666665,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.736328337091045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e8e0fd1bbfb2371f2577","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-588f62a4e071a4f774dcae1b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5361628463309976,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":474,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[236];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":474,"metadata":{"avgScore":0.5361628463309976,"avg_score":0.5361628463309976,"avg_score_percent":53.61628463309977,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":771.62,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":762936773.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":3798417.8333333335,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":93,"runs":301,"runtime":340729.69649999996,"source_flags":[],"source_harness":"codex","source_row_index":236,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":771.62,"totalDurationS":340729.69649999996,"totalInputTokens":762936773.1666666,"totalOutputTokens":3798417.8333333335,"total_cost_usd":771.62,"total_input_tokens":762936773.1666666,"total_output_tokens":3798417.8333333335,"total_runtime_s":340729.69649999996,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.61628463309977,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-83641c02135d1fd77d56","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-51641bccfc664ba3e15ab90d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5094898978576776,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":724,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[361];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":724,"metadata":{"avgScore":0.5094898978576776,"avg_score":0.5094898978576776,"avg_score_percent":50.948989785767765,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":597.25,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":609639112.8333334,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":3521398.8333333335,"passRate":0.26666666666666666,"pass_rate":0.26666666666666666,"pass_rate_percent":26.666666666666668,"passes":54,"runs":211,"runtime":188101.37800000003,"source_flags":[],"source_harness":"codex","source_row_index":361,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":597.25,"totalDurationS":188101.37800000003,"totalInputTokens":609639112.8333334,"totalOutputTokens":3521398.8333333335,"total_cost_usd":597.25,"total_input_tokens":609639112.8333334,"total_output_tokens":3521398.8333333335,"total_runtime_s":188101.37800000003,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.948989785767765,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-21ab494a33b38b63951b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4dba25aacd5b5c1aeb82bd8f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4509751802373446,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":512,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[255];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":512,"metadata":{"avgScore":0.4509751802373446,"avg_score":0.4509751802373446,"avg_score_percent":45.09751802373446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":248.12,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":251962276.66666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1269048.8333333333,"passRate":0.2390350877192982,"pass_rate":0.2390350877192982,"pass_rate_percent":23.903508771929822,"passes":73,"runs":299,"runtime":270652.0256666667,"source_flags":[],"source_harness":"codex","source_row_index":255,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":248.12,"totalDurationS":270652.0256666667,"totalInputTokens":251962276.66666666,"totalOutputTokens":1269048.8333333333,"total_cost_usd":248.12,"total_input_tokens":251962276.66666666,"total_output_tokens":1269048.8333333333,"total_runtime_s":270652.0256666667,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.09751802373446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-52bd998aac2e135ddcbc","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-47597f43c60ae8fb66966cb5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5491172987872457,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":842,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[420];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":842,"metadata":{"avgScore":0.5491172987872457,"avg_score":0.5491172987872457,"avg_score_percent":54.91172987872457,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":525.89,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":543170921.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":2693839,"passRate":0.32978723404255317,"pass_rate":0.32978723404255317,"pass_rate_percent":32.97872340425532,"passes":91,"runs":278,"runtime":282712.47,"source_flags":[],"source_harness":"codex","source_row_index":420,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":525.89,"totalDurationS":282712.47,"totalInputTokens":543170921.5,"totalOutputTokens":2693839,"total_cost_usd":525.89,"total_input_tokens":543170921.5,"total_output_tokens":2693839,"total_runtime_s":282712.47,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.91172987872457,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-622472ce4115a39435f0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-434567d01cf6458b719bb1fd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2746547921284107,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":600,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[299];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":600,"metadata":{"avgScore":0.2746547921284107,"avg_score":0.2746547921284107,"avg_score_percent":27.46547921284107,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":180.82,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":174136662.33333334,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":467479,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":23,"runtime":61723.333333333336,"source_flags":[],"source_harness":"codex","source_row_index":299,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":180.82,"totalDurationS":61723.333333333336,"totalInputTokens":174136662.33333334,"totalOutputTokens":467479,"total_cost_usd":180.82,"total_input_tokens":174136662.33333334,"total_output_tokens":467479,"total_runtime_s":61723.333333333336,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.46547921284107,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d06cfec7d2af7b31879e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3462e8d52cc5045b40d6f30f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.39251001823093684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":160,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[79];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":160,"metadata":{"avgScore":0.39251001823093684,"avg_score":0.39251001823093684,"avg_score_percent":39.25100182309368,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":76.91,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":73118181.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":437364.5,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":22,"runs":107,"runtime":84078.859,"source_flags":[],"source_harness":"codex","source_row_index":79,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":76.91,"totalDurationS":84078.859,"totalInputTokens":73118181.5,"totalOutputTokens":437364.5,"total_cost_usd":76.91,"total_input_tokens":73118181.5,"total_output_tokens":437364.5,"total_runtime_s":84078.859,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.25100182309368,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eec8a152a04dc1a4ae06","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-319ca9d9335d5141c89ebe16","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19372652851069314,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":250,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[124];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":250,"metadata":{"avgScore":0.19372652851069314,"avg_score":0.19372652851069314,"avg_score_percent":19.372652851069315,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":324.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":352326664.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1396392.3333333333,"passRate":0.05263157894736842,"pass_rate":0.05263157894736842,"pass_rate_percent":5.263157894736842,"passes":4,"runs":77,"runtime":139971,"source_flags":[],"source_harness":"codex","source_row_index":124,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":324.61,"totalDurationS":139971,"totalInputTokens":352326664.8333333,"totalOutputTokens":1396392.3333333333,"total_cost_usd":324.61,"total_input_tokens":352326664.8333333,"total_output_tokens":1396392.3333333333,"total_runtime_s":139971,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.372652851069315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-32c866a2806411765372","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-315fcac0c284e1f683f4fa55","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4681714417182331,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":746,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[372];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":746,"metadata":{"avgScore":0.4681714417182331,"avg_score":0.4681714417182331,"avg_score_percent":46.81714417182331,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":132.28,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":140230593.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":820328.5,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":50,"runs":205,"runtime":161026.859,"source_flags":[],"source_harness":"codex","source_row_index":372,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":132.28,"totalDurationS":161026.859,"totalInputTokens":140230593.5,"totalOutputTokens":820328.5,"total_cost_usd":132.28,"total_input_tokens":140230593.5,"total_output_tokens":820328.5,"total_runtime_s":161026.859,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.81714417182331,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2727c2231ece9fceaa02","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-28dd1c20fe271e36c8c4ab41","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4631298337170319,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":960,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[479];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":960,"metadata":{"avgScore":0.4631298337170319,"avg_score":0.4631298337170319,"avg_score_percent":46.31298337170319,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":204.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":177738481.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1150322.5,"passRate":0.3,"pass_rate":0.3,"pass_rate_percent":30.0,"passes":30,"runs":97,"runtime":77278.69649999999,"source_flags":[],"source_harness":"codex","source_row_index":479,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":204.09,"totalDurationS":77278.69649999999,"totalInputTokens":177738481.5,"totalOutputTokens":1150322.5,"total_cost_usd":204.09,"total_input_tokens":177738481.5,"total_output_tokens":1150322.5,"total_runtime_s":77278.69649999999,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.31298337170319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aaccde6f0dd488c8c77f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2035a311ce5767f49ab00325","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.46575439428821086,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":122,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[60];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":122,"metadata":{"avgScore":0.46575439428821086,"avg_score":0.46575439428821086,"avg_score_percent":46.575439428821085,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":256.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":235438015,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1304709,"passRate":0.3,"pass_rate":0.3,"pass_rate_percent":30.0,"passes":33,"runs":106,"runtime":103616.69649999999,"source_flags":[],"source_harness":"codex","source_row_index":60,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":256.09,"totalDurationS":103616.69649999999,"totalInputTokens":235438015,"totalOutputTokens":1304709,"total_cost_usd":256.09,"total_input_tokens":235438015,"total_output_tokens":1304709,"total_runtime_s":103616.69649999999,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.575439428821085,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0127dd0c18a8d58411cf","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1e8cf832f752d7cf50010079","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.16687270502343013,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1090,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[544];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1090,"metadata":{"avgScore":0.16687270502343013,"avg_score":0.16687270502343013,"avg_score_percent":16.68727050234301,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":364.96,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":417867333.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1796854.5,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":2,"runs":61,"runtime":100360.83333333333,"source_flags":[],"source_harness":"codex","source_row_index":544,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":364.96,"totalDurationS":100360.83333333333,"totalInputTokens":417867333.6666667,"totalOutputTokens":1796854.5,"total_cost_usd":364.96,"total_input_tokens":417867333.6666667,"total_output_tokens":1796854.5,"total_runtime_s":100360.83333333333,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.68727050234301,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-27884e16316f265e49b0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1cffe73daaab4c7ce3c0dc29","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5473632833709104,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1316,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[657];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1316,"metadata":{"avgScore":0.5473632833709104,"avg_score":0.5473632833709104,"avg_score_percent":54.736328337091045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":903.28,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":921100178.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":4976045.666666667,"passRate":0.3191489361702128,"pass_rate":0.3191489361702128,"pass_rate_percent":31.914893617021278,"passes":89,"runs":284,"runtime":311864.54466666665,"source_flags":[],"source_harness":"codex","source_row_index":657,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":903.28,"totalDurationS":311864.54466666665,"totalInputTokens":921100178.5,"totalOutputTokens":4976045.666666667,"total_cost_usd":903.28,"total_input_tokens":921100178.5,"total_output_tokens":4976045.666666667,"total_runtime_s":311864.54466666665,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.736328337091045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4000698885b94ac6a7c8","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1aca0a8215634f0a39b1ffa4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19829265389422995,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":630,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[314];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":630,"metadata":{"avgScore":0.19829265389422995,"avg_score":0.19829265389422995,"avg_score_percent":19.829265389422996,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.79,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":50857945.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":196462.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":21,"runtime":57779.5,"source_flags":[],"source_harness":"codex","source_row_index":314,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":49.79,"totalDurationS":57779.5,"totalInputTokens":50857945.5,"totalOutputTokens":196462.5,"total_cost_usd":49.79,"total_input_tokens":50857945.5,"total_output_tokens":196462.5,"total_runtime_s":57779.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.829265389422996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8b8f72f1eda07debe121","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1101e0ece5b7337a9b0c7d34","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44650364852254953,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":964,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[481];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":964,"metadata":{"avgScore":0.44650364852254953,"avg_score":0.44650364852254953,"avg_score_percent":44.65036485225495,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":113.39,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":98748623.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":639137,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":27,"runs":99,"runtime":72734.942,"source_flags":[],"source_harness":"codex","source_row_index":481,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":113.39,"totalDurationS":72734.942,"totalInputTokens":98748623.5,"totalOutputTokens":639137,"total_cost_usd":113.39,"total_input_tokens":98748623.5,"total_output_tokens":639137,"total_runtime_s":72734.942,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.65036485225495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8d1ee16d4798414d22fb","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0d0179cdbce301b6928bb78c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4512408408905345,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":128,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[63];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":128,"metadata":{"avgScore":0.4512408408905345,"avg_score":0.4512408408905345,"avg_score_percent":45.124084089053454,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":298.46,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":286317444,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1756446.1666666667,"passRate":0.2545454545454545,"pass_rate":0.2545454545454545,"pass_rate_percent":25.454545454545453,"passes":27,"runs":109,"runtime":101382.878,"source_flags":[],"source_harness":"codex","source_row_index":63,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":298.46,"totalDurationS":101382.878,"totalInputTokens":286317444,"totalOutputTokens":1756446.1666666667,"total_cost_usd":298.46,"total_input_tokens":286317444,"total_output_tokens":1756446.1666666667,"total_runtime_s":101382.878,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.124084089053454,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-db0a945ada20cf8e87d4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0cdd8e37df52e8e7ab565a1a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4681048909575152,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1348,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[673];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1348,"metadata":{"avgScore":0.4681048909575152,"avg_score":0.4681048909575152,"avg_score_percent":46.81048909575152,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":223.43,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":230074698.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1175061.3333333333,"passRate":0.2541371158392435,"pass_rate":0.2541371158392435,"pass_rate_percent":25.413711583924346,"passes":72,"runs":278,"runtime":242446.52566666668,"source_flags":[],"source_harness":"codex","source_row_index":673,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":223.43,"totalDurationS":242446.52566666668,"totalInputTokens":230074698.16666666,"totalOutputTokens":1175061.3333333333,"total_cost_usd":223.43,"total_input_tokens":230074698.16666666,"total_output_tokens":1175061.3333333333,"total_runtime_s":242446.52566666668,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.81048909575152,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-94bf8dc3fbbdb73b0769","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0508517d0785153b32b3bd5e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.775482331051408,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":360,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[179];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":360,"metadata":{"avgScore":0.775482331051408,"avg_score":0.775482331051408,"avg_score_percent":77.5482331051408,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":172.04,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155266763,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":1008998,"passRate":0.4701492537313433,"pass_rate":0.4701492537313433,"pass_rate_percent":47.01492537313433,"passes":62,"runs":132,"runtime":120620.5,"source_flags":[],"source_harness":"codex","source_row_index":179,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":172.04,"totalDurationS":120620.5,"totalInputTokens":155266763,"totalOutputTokens":1008998,"total_cost_usd":172.04,"total_input_tokens":155266763,"total_output_tokens":1008998,"total_runtime_s":120620.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":77.5482331051408,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7ed0137ccf445d017e99","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-04e28f93128e4a300449ede9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19420078029613763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c610bca43afe9d6d0751c9d63e8806eba16a484","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":248,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[123];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":248,"metadata":{"avgScore":0.19420078029613763,"avg_score":0.19420078029613763,"avg_score_percent":19.420078029613762,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":249.88,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":294882387,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Sol","output_tokens":783893.5,"passRate":0.039473684210526314,"pass_rate":0.039473684210526314,"pass_rate_percent":3.9473684210526314,"passes":2,"runs":75,"runtime":117354.5,"source_flags":[],"source_harness":"codex","source_row_index":123,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":249.88,"totalDurationS":117354.5,"totalInputTokens":294882387,"totalOutputTokens":783893.5,"total_cost_usd":249.88,"total_input_tokens":294882387,"total_output_tokens":783893.5,"total_runtime_s":117354.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.420078029613762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-09ec73300557df3a9222","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f895764d20cd3a330716ef45","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2857142857142857,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":711,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[355];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":711,"metadata":{"avgScore":0.531374869029037,"avg_score":0.531374869029037,"avg_score_percent":53.137486902903696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":279.31,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":300085777.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1422241,"passRate":0.2857142857142857,"pass_rate":0.2857142857142857,"pass_rate_percent":28.57142857142857,"passes":57,"runs":207,"runtime":171176.44199999998,"source_flags":[],"source_harness":"codex","source_row_index":355,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":279.31,"totalDurationS":171176.44199999998,"totalInputTokens":300085777.5,"totalOutputTokens":1422241,"total_cost_usd":279.31,"total_input_tokens":300085777.5,"total_output_tokens":1422241,"total_runtime_s":171176.44199999998,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.57142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fba49951729c9f353cf7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f62e026202d8d2ee4595bc52","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2857142857142857,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":709,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[354];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":709,"metadata":{"avgScore":0.5363939737297058,"avg_score":0.5363939737297058,"avg_score_percent":53.63939737297058,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446.38,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":443472815.3333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2515905.5,"passRate":0.2857142857142857,"pass_rate":0.2857142857142857,"pass_rate_percent":28.57142857142857,"passes":60,"runs":209,"runtime":179145.1965,"source_flags":[],"source_harness":"codex","source_row_index":354,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":446.38,"totalDurationS":179145.1965,"totalInputTokens":443472815.3333333,"totalOutputTokens":2515905.5,"total_cost_usd":446.38,"total_input_tokens":443472815.3333333,"total_output_tokens":2515905.5,"total_runtime_s":179145.1965,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.57142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bf8d5772a39a54571d06","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f32d75b16f36cb62f65055d7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3191489361702128,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":843,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[421];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":843,"metadata":{"avgScore":0.5473632833709104,"avg_score":0.5473632833709104,"avg_score_percent":54.736328337091045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":903.28,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":921100178.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":4976045.666666667,"passRate":0.3191489361702128,"pass_rate":0.3191489361702128,"pass_rate_percent":31.914893617021278,"passes":89,"runs":284,"runtime":311864.54466666665,"source_flags":[],"source_harness":"codex","source_row_index":421,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":903.28,"totalDurationS":311864.54466666665,"totalInputTokens":921100178.5,"totalOutputTokens":4976045.666666667,"total_cost_usd":903.28,"total_input_tokens":921100178.5,"total_output_tokens":4976045.666666667,"total_runtime_s":311864.54466666665,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.914893617021278,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-efde0356be6905f95a90","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ec5cf09f71b2354158ce7278","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1089,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[544];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1089,"metadata":{"avgScore":0.16687270502343013,"avg_score":0.16687270502343013,"avg_score_percent":16.68727050234301,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":364.96,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":417867333.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1796854.5,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":2,"runs":61,"runtime":100360.83333333333,"source_flags":[],"source_harness":"codex","source_row_index":544,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":364.96,"totalDurationS":100360.83333333333,"totalInputTokens":417867333.6666667,"totalOutputTokens":1796854.5,"total_cost_usd":364.96,"total_input_tokens":417867333.6666667,"total_output_tokens":1796854.5,"total_runtime_s":100360.83333333333,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ea6f6e2e1dd9a26d436d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e847e60c1d6498af87c4be41","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4552238805970149,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":357,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[178];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":357,"metadata":{"avgScore":0.7829241382323966,"avg_score":0.7829241382323966,"avg_score_percent":78.29241382323966,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":210.38,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":189978516.33333334,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1282200,"passRate":0.4552238805970149,"pass_rate":0.4552238805970149,"pass_rate_percent":45.52238805970149,"passes":61,"runs":134,"runtime":125006.5,"source_flags":[],"source_harness":"codex","source_row_index":178,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":210.38,"totalDurationS":125006.5,"totalInputTokens":189978516.33333334,"totalOutputTokens":1282200,"total_cost_usd":210.38,"total_input_tokens":189978516.33333334,"total_output_tokens":1282200,"total_runtime_s":125006.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.52238805970149,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-06c292ccdaaeb7f6bfc3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e4f0a1356a56d8227f46b759","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":963,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[481];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":963,"metadata":{"avgScore":0.44650364852254953,"avg_score":0.44650364852254953,"avg_score_percent":44.65036485225495,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":113.39,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":98748623.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":639137,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":27,"runs":99,"runtime":72734.942,"source_flags":[],"source_harness":"codex","source_row_index":481,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":113.39,"totalDurationS":72734.942,"totalInputTokens":98748623.5,"totalOutputTokens":639137,"total_cost_usd":113.39,"total_input_tokens":98748623.5,"total_output_tokens":639137,"total_runtime_s":72734.942,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-54afb560d2e651796c27","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e3707e6372a46c838b5e08f6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3059210526315789,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":473,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[236];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":473,"metadata":{"avgScore":0.5361628463309976,"avg_score":0.5361628463309976,"avg_score_percent":53.61628463309977,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":771.62,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":762936773.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":3798417.8333333335,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":93,"runs":301,"runtime":340729.69649999996,"source_flags":[],"source_harness":"codex","source_row_index":236,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":771.62,"totalDurationS":340729.69649999996,"totalInputTokens":762936773.1666666,"totalOutputTokens":3798417.8333333335,"total_cost_usd":771.62,"total_input_tokens":762936773.1666666,"total_output_tokens":3798417.8333333335,"total_runtime_s":340729.69649999996,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.592105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8f87df250ee13a6d310","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-da24ea978abf610ff146a204","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.32978723404255317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1313,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[656];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1313,"metadata":{"avgScore":0.5491172987872457,"avg_score":0.5491172987872457,"avg_score_percent":54.91172987872457,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":525.89,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":543170921.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2693839,"passRate":0.32978723404255317,"pass_rate":0.32978723404255317,"pass_rate_percent":32.97872340425532,"passes":91,"runs":278,"runtime":282712.47,"source_flags":[],"source_harness":"codex","source_row_index":656,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":525.89,"totalDurationS":282712.47,"totalInputTokens":543170921.5,"totalOutputTokens":2693839,"total_cost_usd":525.89,"total_input_tokens":543170921.5,"total_output_tokens":2693839,"total_runtime_s":282712.47,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.97872340425532,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-03a9356893cbe64d8ff6","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d479de8e794b83ae7595adfd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2541371158392435,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":875,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[437];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":875,"metadata":{"avgScore":0.4681048909575152,"avg_score":0.4681048909575152,"avg_score_percent":46.81048909575152,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":223.43,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":230074698.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1175061.3333333333,"passRate":0.2541371158392435,"pass_rate":0.2541371158392435,"pass_rate_percent":25.413711583924346,"passes":72,"runs":278,"runtime":242446.52566666668,"source_flags":[],"source_harness":"codex","source_row_index":437,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":223.43,"totalDurationS":242446.52566666668,"totalInputTokens":230074698.16666666,"totalOutputTokens":1175061.3333333333,"total_cost_usd":223.43,"total_input_tokens":230074698.16666666,"total_output_tokens":1175061.3333333333,"total_runtime_s":242446.52566666668,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.413711583924346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ff6418605fd95c80af6a","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d1bd3444c747bd3e34b376b4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.06666666666666667,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1079,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[539];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1079,"metadata":{"avgScore":0.20737264567839306,"avg_score":0.20737264567839306,"avg_score_percent":20.737264567839304,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":259.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":297768333.3333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1208400.8333333333,"passRate":0.06666666666666667,"pass_rate":0.06666666666666667,"pass_rate_percent":6.666666666666667,"passes":4,"runs":61,"runtime":97842,"source_flags":[],"source_harness":"codex","source_row_index":539,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":259.53,"totalDurationS":97842,"totalInputTokens":297768333.3333333,"totalOutputTokens":1208400.8333333333,"total_cost_usd":259.53,"total_input_tokens":297768333.3333333,"total_output_tokens":1208400.8333333333,"total_runtime_s":97842,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.666666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5f30def9c2a018b84aa4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cfa5c3d4bd3a2d7abab5c2d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2545454545454545,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":127,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[63];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":127,"metadata":{"avgScore":0.4512408408905345,"avg_score":0.4512408408905345,"avg_score_percent":45.124084089053454,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":298.46,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":286317444,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1756446.1666666667,"passRate":0.2545454545454545,"pass_rate":0.2545454545454545,"pass_rate_percent":25.454545454545453,"passes":27,"runs":109,"runtime":101382.878,"source_flags":[],"source_harness":"codex","source_row_index":63,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":298.46,"totalDurationS":101382.878,"totalInputTokens":286317444,"totalOutputTokens":1756446.1666666667,"total_cost_usd":298.46,"total_input_tokens":286317444,"total_output_tokens":1756446.1666666667,"total_runtime_s":101382.878,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.454545454545453,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5d4eeafdcadfd25caf2b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ce6470e9ed5c7291d5d8b79c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3191489361702128,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1315,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[657];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1315,"metadata":{"avgScore":0.5473632833709104,"avg_score":0.5473632833709104,"avg_score_percent":54.736328337091045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":903.28,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":921100178.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":4976045.666666667,"passRate":0.3191489361702128,"pass_rate":0.3191489361702128,"pass_rate_percent":31.914893617021278,"passes":89,"runs":284,"runtime":311864.54466666665,"source_flags":[],"source_harness":"codex","source_row_index":657,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":903.28,"totalDurationS":311864.54466666665,"totalInputTokens":921100178.5,"totalOutputTokens":4976045.666666667,"total_cost_usd":903.28,"total_input_tokens":921100178.5,"total_output_tokens":4976045.666666667,"total_runtime_s":311864.54466666665,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.914893617021278,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3de6716b3049ebf89d69","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c81b11be284c735507c4ef16","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29495614035087714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":481,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[240];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":481,"metadata":{"avgScore":0.5210026942935828,"avg_score":0.5210026942935828,"avg_score_percent":52.10026942935828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":518.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":539877934.6666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2196349.5,"passRate":0.29495614035087714,"pass_rate":0.29495614035087714,"pass_rate_percent":29.495614035087712,"passes":87,"runs":302,"runtime":328693.94200000004,"source_flags":[],"source_harness":"codex","source_row_index":240,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":518.72,"totalDurationS":328693.94200000004,"totalInputTokens":539877934.6666666,"totalOutputTokens":2196349.5,"total_cost_usd":518.72,"total_input_tokens":539877934.6666666,"total_output_tokens":2196349.5,"total_runtime_s":328693.94200000004,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.495614035087712,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8afca266fa406af14b48","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c498b38a1ce4237bd457f515","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29605263157894735,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":3,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[1];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":3,"metadata":{"avgScore":0.5276278004520454,"avg_score":0.5276278004520454,"avg_score_percent":52.762780045204536,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1084.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1095236840.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":5443524.666666667,"passRate":0.29605263157894735,"pass_rate":0.29605263157894735,"pass_rate_percent":29.605263157894733,"passes":89,"runs":307,"runtime":373587.878,"source_flags":[],"source_harness":"codex","source_row_index":1,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":1084.09,"totalDurationS":373587.878,"totalInputTokens":1095236840.8333333,"totalOutputTokens":5443524.666666667,"total_cost_usd":1084.09,"total_input_tokens":1095236840.8333333,"total_output_tokens":5443524.666666667,"total_runtime_s":373587.878,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.605263157894733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a30e9690b96429a4b394","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c33b188fe171f76d50fd30b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":159,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[79];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":159,"metadata":{"avgScore":0.39251001823093684,"avg_score":0.39251001823093684,"avg_score_percent":39.25100182309368,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":76.91,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":73118181.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":437364.5,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":22,"runs":107,"runtime":84078.859,"source_flags":[],"source_harness":"codex","source_row_index":79,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":76.91,"totalDurationS":84078.859,"totalInputTokens":73118181.5,"totalOutputTokens":437364.5,"total_cost_usd":76.91,"total_input_tokens":73118181.5,"total_output_tokens":437364.5,"total_runtime_s":84078.859,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e4633eab5aba5c960d97","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c27705a3f3925820fafbdd40","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2714285714285714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":719,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[359];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":719,"metadata":{"avgScore":0.512155695633506,"avg_score":0.512155695633506,"avg_score_percent":51.215569563350606,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":343.41,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":359092352,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1936844.5,"passRate":0.2714285714285714,"pass_rate":0.2714285714285714,"pass_rate_percent":27.142857142857142,"passes":55,"runs":206,"runtime":184402.47,"source_flags":[],"source_harness":"codex","source_row_index":359,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":343.41,"totalDurationS":184402.47,"totalInputTokens":359092352,"totalOutputTokens":1936844.5,"total_cost_usd":343.41,"total_input_tokens":359092352,"total_output_tokens":1936844.5,"total_runtime_s":184402.47,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.142857142857142,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-62bcfc5fa5508aef0572","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bb767c2e3231778adad6f5f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.47761194029850745,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":355,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[177];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":355,"metadata":{"avgScore":0.7882473228071811,"avg_score":0.7882473228071811,"avg_score_percent":78.82473228071811,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":322.09,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":304163441.3333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1816516,"passRate":0.47761194029850745,"pass_rate":0.47761194029850745,"pass_rate_percent":47.76119402985074,"passes":63,"runs":134,"runtime":145737.33333333334,"source_flags":[],"source_harness":"codex","source_row_index":177,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":322.09,"totalDurationS":145737.33333333334,"totalInputTokens":304163441.3333333,"totalOutputTokens":1816516,"total_cost_usd":322.09,"total_input_tokens":304163441.3333333,"total_output_tokens":1816516,"total_runtime_s":145737.33333333334,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.76119402985074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a302546173ae1a2c4cb8","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b34da02cd7cb5a7c92e4c583","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3059210526315789,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[0];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1,"metadata":{"avgScore":0.5361628463309976,"avg_score":0.5361628463309976,"avg_score_percent":53.61628463309977,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":771.62,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":762936773.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":3798417.8333333335,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":93,"runs":301,"runtime":340729.69649999996,"source_flags":[],"source_harness":"codex","source_row_index":0,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":771.62,"totalDurationS":340729.69649999996,"totalInputTokens":762936773.1666666,"totalOutputTokens":3798417.8333333335,"total_cost_usd":771.62,"total_input_tokens":762936773.1666666,"total_output_tokens":3798417.8333333335,"total_runtime_s":340729.69649999996,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.592105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-48b14195a76d8ab1b220","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-afe219a3a6bc4ba9c74ee4fa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29605263157894735,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":475,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[237];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":475,"metadata":{"avgScore":0.5276278004520454,"avg_score":0.5276278004520454,"avg_score_percent":52.762780045204536,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1084.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1095236840.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":5443524.666666667,"passRate":0.29605263157894735,"pass_rate":0.29605263157894735,"pass_rate_percent":29.605263157894733,"passes":89,"runs":307,"runtime":373587.878,"source_flags":[],"source_harness":"codex","source_row_index":237,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":1084.09,"totalDurationS":373587.878,"totalInputTokens":1095236840.8333333,"totalOutputTokens":5443524.666666667,"total_cost_usd":1084.09,"total_input_tokens":1095236840.8333333,"total_output_tokens":5443524.666666667,"total_runtime_s":373587.878,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.605263157894733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-75030ec02d3aedd6ce0b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a6af3f1f744af718c9c8ddd0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.32978723404255317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":841,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[420];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":841,"metadata":{"avgScore":0.5491172987872457,"avg_score":0.5491172987872457,"avg_score_percent":54.91172987872457,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":525.89,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":543170921.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2693839,"passRate":0.32978723404255317,"pass_rate":0.32978723404255317,"pass_rate_percent":32.97872340425532,"passes":91,"runs":278,"runtime":282712.47,"source_flags":[],"source_harness":"codex","source_row_index":420,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":525.89,"totalDurationS":282712.47,"totalInputTokens":543170921.5,"totalOutputTokens":2693839,"total_cost_usd":525.89,"total_input_tokens":543170921.5,"total_output_tokens":2693839,"total_runtime_s":282712.47,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.97872340425532,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-536c18f3984a4b4dbd97","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-92c4b8e100b8420aa75711a3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2727272727272727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":139,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[69];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":139,"metadata":{"avgScore":0.4301432864776319,"avg_score":0.4301432864776319,"avg_score_percent":43.014328647763186,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":164.83,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":153018341,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":942915.5,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":30,"runs":108,"runtime":107709.47,"source_flags":[],"source_harness":"codex","source_row_index":69,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":164.83,"totalDurationS":107709.47,"totalInputTokens":153018341,"totalOutputTokens":942915.5,"total_cost_usd":164.83,"total_input_tokens":153018341,"total_output_tokens":942915.5,"total_runtime_s":107709.47,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c5bace9eacc381a3b632","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-929f1a848d0d2a260e4b821a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":973,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[486];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":973,"metadata":{"avgScore":0.433638270435078,"avg_score":0.433638270435078,"avg_score_percent":43.363827043507804,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":138.39,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":124762948,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":846635.5,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":28,"runs":99,"runtime":81367.47,"source_flags":[],"source_harness":"codex","source_row_index":486,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":138.39,"totalDurationS":81367.47,"totalInputTokens":124762948,"totalOutputTokens":846635.5,"total_cost_usd":138.39,"total_input_tokens":124762948,"total_output_tokens":846635.5,"total_runtime_s":81367.47,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9003f1984272a2347dd7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8bdf7df57274bdd28b1b5006","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29495614035087714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":9,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[4];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":9,"metadata":{"avgScore":0.5210026942935828,"avg_score":0.5210026942935828,"avg_score_percent":52.10026942935828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":518.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":539877934.6666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2196349.5,"passRate":0.29495614035087714,"pass_rate":0.29495614035087714,"pass_rate_percent":29.495614035087712,"passes":87,"runs":302,"runtime":328693.94200000004,"source_flags":[],"source_harness":"codex","source_row_index":4,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":518.72,"totalDurationS":328693.94200000004,"totalInputTokens":539877934.6666666,"totalOutputTokens":2196349.5,"total_cost_usd":518.72,"total_input_tokens":539877934.6666666,"total_output_tokens":2196349.5,"total_runtime_s":328693.94200000004,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.495614035087712,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f4b80e9672fef08ea802","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-893f3988262f4489e52b8bed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":997,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[498];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":997,"metadata":{"avgScore":0.39285408160616503,"avg_score":0.39285408160616503,"avg_score_percent":39.285408160616505,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.9,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":59225023.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":389348.5,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":20,"runs":98,"runtime":65719.859,"source_flags":[],"source_harness":"codex","source_row_index":498,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":61.9,"totalDurationS":65719.859,"totalInputTokens":59225023.5,"totalOutputTokens":389348.5,"total_cost_usd":61.9,"total_input_tokens":59225023.5,"total_output_tokens":389348.5,"total_runtime_s":65719.859,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9515fd2c4be351f4973c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-842de8e3747a7a59b2ae7e44","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4701492537313433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":359,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[179];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":359,"metadata":{"avgScore":0.775482331051408,"avg_score":0.775482331051408,"avg_score_percent":77.5482331051408,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":172.04,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155266763,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1008998,"passRate":0.4701492537313433,"pass_rate":0.4701492537313433,"pass_rate_percent":47.01492537313433,"passes":62,"runs":132,"runtime":120620.5,"source_flags":[],"source_harness":"codex","source_row_index":179,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":172.04,"totalDurationS":120620.5,"totalInputTokens":155266763,"totalOutputTokens":1008998,"total_cost_usd":172.04,"total_input_tokens":155266763,"total_output_tokens":1008998,"total_runtime_s":120620.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.01492537313433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c35f5e3e742b168add07","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-816d18872b7d8aadc73ed773","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4701492537313433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1195,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[597];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1195,"metadata":{"avgScore":0.775482331051408,"avg_score":0.775482331051408,"avg_score_percent":77.5482331051408,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":172.04,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155266763,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1008998,"passRate":0.4701492537313433,"pass_rate":0.4701492537313433,"pass_rate_percent":47.01492537313433,"passes":62,"runs":132,"runtime":120620.5,"source_flags":[],"source_harness":"codex","source_row_index":597,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":172.04,"totalDurationS":120620.5,"totalInputTokens":155266763,"totalOutputTokens":1008998,"total_cost_usd":172.04,"total_input_tokens":155266763,"total_output_tokens":1008998,"total_runtime_s":120620.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.01492537313433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-59da6d1eab4953a59386","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-80d594a868bf9a1177ee8ce3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":605,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[302];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":605,"metadata":{"avgScore":0.23849458053696015,"avg_score":0.23849458053696015,"avg_score_percent":23.849458053696015,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":94.47,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":92690197.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":190449,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":55159.5,"source_flags":[],"source_harness":"codex","source_row_index":302,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":94.47,"totalDurationS":55159.5,"totalInputTokens":92690197.5,"totalOutputTokens":190449,"total_cost_usd":94.47,"total_input_tokens":92690197.5,"total_output_tokens":190449,"total_runtime_s":55159.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8dac8df97dc49cc478fd","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7cf18b925909ef5a66788a31","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":275,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[137];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":275,"metadata":{"avgScore":0.1261806657735013,"avg_score":0.1261806657735013,"avg_score_percent":12.61806657735013,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":96.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":111296708,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":408600.3333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":74,"runtime":109927.33333333333,"source_flags":[],"source_harness":"codex","source_row_index":137,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":96.13,"totalDurationS":109927.33333333333,"totalInputTokens":111296708,"totalOutputTokens":408600.3333333333,"total_cost_usd":96.13,"total_input_tokens":111296708,"total_output_tokens":408600.3333333333,"total_runtime_s":109927.33333333333,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-169c1d9804c14a09d259","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7b4e3b3bd6663c9f46d82ba6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26666666666666666,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":723,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[361];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":723,"metadata":{"avgScore":0.5094898978576776,"avg_score":0.5094898978576776,"avg_score_percent":50.948989785767765,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":597.25,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":609639112.8333334,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":3521398.8333333335,"passRate":0.26666666666666666,"pass_rate":0.26666666666666666,"pass_rate_percent":26.666666666666668,"passes":54,"runs":211,"runtime":188101.37800000003,"source_flags":[],"source_harness":"codex","source_row_index":361,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":597.25,"totalDurationS":188101.37800000003,"totalInputTokens":609639112.8333334,"totalOutputTokens":3521398.8333333335,"total_cost_usd":597.25,"total_input_tokens":609639112.8333334,"total_output_tokens":3521398.8333333335,"total_runtime_s":188101.37800000003,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.666666666666668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c9f0cfb67b6729d27ef2","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-79dab6d53da10a079137d60f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.47761194029850745,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1191,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[595];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1191,"metadata":{"avgScore":0.7882473228071811,"avg_score":0.7882473228071811,"avg_score_percent":78.82473228071811,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":322.09,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":304163441.3333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1816516,"passRate":0.47761194029850745,"pass_rate":0.47761194029850745,"pass_rate_percent":47.76119402985074,"passes":63,"runs":134,"runtime":145737.33333333334,"source_flags":[],"source_harness":"codex","source_row_index":595,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":322.09,"totalDurationS":145737.33333333334,"totalInputTokens":304163441.3333333,"totalOutputTokens":1816516,"total_cost_usd":322.09,"total_input_tokens":304163441.3333333,"total_output_tokens":1816516,"total_runtime_s":145737.33333333334,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.76119402985074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1ddc2488822ecf59749a","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-759440e050abaea460ce9b63","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3059210526315789,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":7,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[3];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":7,"metadata":{"avgScore":0.523728673169988,"avg_score":0.523728673169988,"avg_score_percent":52.37286731699879,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":575.68,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":594028867,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2890301.5,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":91,"runs":299,"runtime":340491.97,"source_flags":[],"source_harness":"codex","source_row_index":3,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":575.68,"totalDurationS":340491.97,"totalInputTokens":594028867,"totalOutputTokens":2890301.5,"total_cost_usd":575.68,"total_input_tokens":594028867,"total_output_tokens":2890301.5,"total_runtime_s":340491.97,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.592105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5104f8e6816f8785e66d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-758c55b3e6da0aeed4e52b8f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3059210526315789,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":479,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[239];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":479,"metadata":{"avgScore":0.523728673169988,"avg_score":0.523728673169988,"avg_score_percent":52.37286731699879,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":575.68,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":594028867,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2890301.5,"passRate":0.3059210526315789,"pass_rate":0.3059210526315789,"pass_rate_percent":30.592105263157894,"passes":91,"runs":299,"runtime":340491.97,"source_flags":[],"source_harness":"codex","source_row_index":239,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":575.68,"totalDurationS":340491.97,"totalInputTokens":594028867,"totalOutputTokens":2890301.5,"total_cost_usd":575.68,"total_input_tokens":594028867,"total_output_tokens":2890301.5,"total_runtime_s":340491.97,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.592105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8f78c15f129852cda95","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6c77f44ede347361bbec808a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":629,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[314];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":629,"metadata":{"avgScore":0.19829265389422995,"avg_score":0.19829265389422995,"avg_score_percent":19.829265389422996,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.79,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":50857945.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":196462.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":21,"runtime":57779.5,"source_flags":[],"source_harness":"codex","source_row_index":314,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":49.79,"totalDurationS":57779.5,"totalInputTokens":50857945.5,"totalOutputTokens":196462.5,"total_cost_usd":49.79,"total_input_tokens":50857945.5,"total_output_tokens":196462.5,"total_runtime_s":57779.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-02294eb8026a902a7690","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-698ac221ece4a0dfbb02dfc1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":255,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[127];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":255,"metadata":{"avgScore":0.1644048385293532,"avg_score":0.1644048385293532,"avg_score_percent":16.44048385293532,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":488.13,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":523793885.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2091747,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":2,"runs":79,"runtime":154166.1666666667,"source_flags":[],"source_harness":"codex","source_row_index":127,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":488.13,"totalDurationS":154166.1666666667,"totalInputTokens":523793885.5,"totalOutputTokens":2091747,"total_cost_usd":488.13,"total_input_tokens":523793885.5,"total_output_tokens":2091747,"total_runtime_s":154166.1666666667,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-060bd889f2b803861701","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6762c9486816db6ce1a3d93a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3930348258706468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":395,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[197];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":395,"metadata":{"avgScore":0.6677679137696316,"avg_score":0.6677679137696316,"avg_score_percent":66.77679137696316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.16,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":71044249.66666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":467625,"passRate":0.3930348258706468,"pass_rate":0.3930348258706468,"pass_rate_percent":39.30348258706468,"passes":53,"runs":134,"runtime":96179.83333333334,"source_flags":[],"source_harness":"codex","source_row_index":197,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":80.16,"totalDurationS":96179.83333333334,"totalInputTokens":71044249.66666667,"totalOutputTokens":467625,"total_cost_usd":80.16,"total_input_tokens":71044249.66666667,"total_output_tokens":467625,"total_runtime_s":96179.83333333334,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.30348258706468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1fdb902abc9267d172a","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-63c70b76f6a3e807fbc8c612","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2390350877192982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":511,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[255];split=full/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":511,"metadata":{"avgScore":0.4509751802373446,"avg_score":0.4509751802373446,"avg_score_percent":45.09751802373446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":248.12,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":251962276.66666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1269048.8333333333,"passRate":0.2390350877192982,"pass_rate":0.2390350877192982,"pass_rate_percent":23.903508771929822,"passes":73,"runs":299,"runtime":270652.0256666667,"source_flags":[],"source_harness":"codex","source_row_index":255,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":248.12,"totalDurationS":270652.0256666667,"totalInputTokens":251962276.66666666,"totalOutputTokens":1269048.8333333333,"total_cost_usd":248.12,"total_input_tokens":251962276.66666666,"total_output_tokens":1269048.8333333333,"total_runtime_s":270652.0256666667,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.903508771929822,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-329ff530848dd19ee9a0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5fc96c43fa883dcd493c4701","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":959,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[479];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":959,"metadata":{"avgScore":0.4631298337170319,"avg_score":0.4631298337170319,"avg_score_percent":46.31298337170319,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":204.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":177738481.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1150322.5,"passRate":0.3,"pass_rate":0.3,"pass_rate_percent":30.0,"passes":30,"runs":97,"runtime":77278.69649999999,"source_flags":[],"source_harness":"codex","source_row_index":479,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":204.09,"totalDurationS":77278.69649999999,"totalInputTokens":177738481.5,"totalOutputTokens":1150322.5,"total_cost_usd":204.09,"total_input_tokens":177738481.5,"total_output_tokens":1150322.5,"total_runtime_s":77278.69649999999,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8178e8eb83ea76c8a5dc","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-55c1725f62054def4bc37ab8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":121,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[60];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":121,"metadata":{"avgScore":0.46575439428821086,"avg_score":0.46575439428821086,"avg_score_percent":46.575439428821085,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":256.09,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":235438015,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1304709,"passRate":0.3,"pass_rate":0.3,"pass_rate_percent":30.0,"passes":33,"runs":106,"runtime":103616.69649999999,"source_flags":[],"source_harness":"codex","source_row_index":60,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":256.09,"totalDurationS":103616.69649999999,"totalInputTokens":235438015,"totalOutputTokens":1304709,"total_cost_usd":256.09,"total_input_tokens":235438015,"total_output_tokens":1304709,"total_runtime_s":103616.69649999999,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6c524805a8495a325fc8","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-471da0f64a0d655c030676f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.45273631840796025,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1199,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[599];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1199,"metadata":{"avgScore":0.7561914737576828,"avg_score":0.7561914737576828,"avg_score_percent":75.61914737576828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":136.75,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":124228619.66666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":787503.5,"passRate":0.45273631840796025,"pass_rate":0.45273631840796025,"pass_rate_percent":45.27363184079603,"passes":60,"runs":134,"runtime":123361.5,"source_flags":[],"source_harness":"codex","source_row_index":599,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":136.75,"totalDurationS":123361.5,"totalInputTokens":124228619.66666667,"totalOutputTokens":787503.5,"total_cost_usd":136.75,"total_input_tokens":124228619.66666667,"total_output_tokens":787503.5,"total_runtime_s":123361.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.27363184079603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b39c41c3756397e0a0d4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-46ffd099111ea9e726bd9dca","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.06666666666666667,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1085,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[542];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1085,"metadata":{"avgScore":0.19652298994758152,"avg_score":0.19652298994758152,"avg_score_percent":19.65229899475815,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":227.28,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":273176548.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":948762,"passRate":0.06666666666666667,"pass_rate":0.06666666666666667,"pass_rate_percent":6.666666666666667,"passes":3,"runs":59,"runtime":89607,"source_flags":[],"source_harness":"codex","source_row_index":542,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":227.28,"totalDurationS":89607,"totalInputTokens":273176548.5,"totalOutputTokens":948762,"total_cost_usd":227.28,"total_input_tokens":273176548.5,"total_output_tokens":948762,"total_runtime_s":89607,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.666666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4b6d9532ccf49bcabda1","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3e6794a94397cd4524f8a75a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.317966903073286,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":845,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[422];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":845,"metadata":{"avgScore":0.5430423343738866,"avg_score":0.5430423343738866,"avg_score_percent":54.30423343738866,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":424.24,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":447187737.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2005900.5,"passRate":0.317966903073286,"pass_rate":0.317966903073286,"pass_rate_percent":31.7966903073286,"passes":87,"runs":280,"runtime":273534.44200000004,"source_flags":[],"source_harness":"codex","source_row_index":422,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":424.24,"totalDurationS":273534.44200000004,"totalInputTokens":447187737.1666667,"totalOutputTokens":2005900.5,"total_cost_usd":424.24,"total_input_tokens":447187737.1666667,"total_output_tokens":2005900.5,"total_runtime_s":273534.44200000004,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.7966903073286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0dffad978ac44b6d29ec","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3d6f9c1ff388ab470bf83579","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.05263157894736842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":249,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[124];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":249,"metadata":{"avgScore":0.19372652851069314,"avg_score":0.19372652851069314,"avg_score_percent":19.372652851069315,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":324.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":352326664.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1396392.3333333333,"passRate":0.05263157894736842,"pass_rate":0.05263157894736842,"pass_rate_percent":5.263157894736842,"passes":4,"runs":77,"runtime":139971,"source_flags":[],"source_harness":"codex","source_row_index":124,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":324.61,"totalDurationS":139971,"totalInputTokens":352326664.8333333,"totalOutputTokens":1396392.3333333333,"total_cost_usd":324.61,"total_input_tokens":352326664.8333333,"total_output_tokens":1396392.3333333333,"total_runtime_s":139971,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":5.263157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-da7ea35defdfa778a74c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3d545d6a9d2bc0590d6f94c2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3262411347517731,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1309,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[654];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1309,"metadata":{"avgScore":0.5595483966613978,"avg_score":0.5595483966613978,"avg_score_percent":55.954839666139776,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":656.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":652364022.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":3478380.3333333335,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":92,"runs":280,"runtime":291953.19649999996,"source_flags":[],"source_harness":"codex","source_row_index":654,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":656.72,"totalDurationS":291953.19649999996,"totalInputTokens":652364022.1666666,"totalOutputTokens":3478380.3333333335,"total_cost_usd":656.72,"total_input_tokens":652364022.1666666,"total_output_tokens":3478380.3333333335,"total_runtime_s":291953.19649999996,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.62411347517731,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a7b8e745d0abaefe6f9b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3c412d2b144354e8146470ea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":967,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[483];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":967,"metadata":{"avgScore":0.44076492497958797,"avg_score":0.44076492497958797,"avg_score_percent":44.076492497958796,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":239.3,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":217198822,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1560829.1666666667,"passRate":0.26,"pass_rate":0.26,"pass_rate_percent":26.0,"passes":25,"runs":100,"runtime":74690.378,"source_flags":[],"source_harness":"codex","source_row_index":483,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":239.3,"totalDurationS":74690.378,"totalInputTokens":217198822,"totalOutputTokens":1560829.1666666667,"total_cost_usd":239.3,"total_input_tokens":217198822,"total_output_tokens":1560829.1666666667,"total_runtime_s":74690.378,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8be8eb01bf0279c98bf3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3aaafbe36f9219a52c74979a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.045454545454545456,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":607,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[303];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":607,"metadata":{"avgScore":0.23640261027768616,"avg_score":0.23640261027768616,"avg_score_percent":23.640261027768616,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":114.9,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":110572751,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":320037.5,"passRate":0.045454545454545456,"pass_rate":0.045454545454545456,"pass_rate_percent":4.545454545454546,"passes":1,"runs":21,"runtime":48776.5,"source_flags":[],"source_harness":"codex","source_row_index":303,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":114.9,"totalDurationS":48776.5,"totalInputTokens":110572751,"totalOutputTokens":320037.5,"total_cost_usd":114.9,"total_input_tokens":110572751,"total_output_tokens":320037.5,"total_runtime_s":48776.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":4.545454545454546,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ee20a55005878cd858a3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-39fe5f3894166737d64e201a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2727272727272727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":133,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[66];split=full/full-spectrum;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":133,"metadata":{"avgScore":0.4451610959401981,"avg_score":0.4451610959401981,"avg_score_percent":44.51610959401981,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":157.67,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":142119530.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":732332.5,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":29,"runs":109,"runtime":107957.442,"source_flags":[],"source_harness":"codex","source_row_index":66,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":157.67,"totalDurationS":107957.442,"totalInputTokens":142119530.5,"totalOutputTokens":732332.5,"total_cost_usd":157.67,"total_input_tokens":142119530.5,"total_output_tokens":732332.5,"total_runtime_s":107957.442,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-117bed5c013ddf1f2c60","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3764df512e1ab415f3f4940d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.05,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1087,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[543];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1087,"metadata":{"avgScore":0.18922557053100295,"avg_score":0.18922557053100295,"avg_score_percent":18.922557053100295,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":183.03,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":231236782,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":657382,"passRate":0.05,"pass_rate":0.05,"pass_rate_percent":5.0,"passes":2,"runs":59,"runtime":84648,"source_flags":[],"source_harness":"codex","source_row_index":543,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":183.03,"totalDurationS":84648,"totalInputTokens":231236782,"totalOutputTokens":657382,"total_cost_usd":183.03,"total_input_tokens":231236782,"total_output_tokens":657382,"total_runtime_s":84648,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":5.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-df26442e501f652167cf","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-342cad2e8ef529708e026ec8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":599,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[299];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":599,"metadata":{"avgScore":0.2746547921284107,"avg_score":0.2746547921284107,"avg_score_percent":27.46547921284107,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":180.82,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":174136662.33333334,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":467479,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":23,"runtime":61723.333333333336,"source_flags":[],"source_harness":"codex","source_row_index":299,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":180.82,"totalDurationS":61723.333333333336,"totalInputTokens":174136662.33333334,"totalOutputTokens":467479,"total_cost_usd":180.82,"total_input_tokens":174136662.33333334,"total_output_tokens":467479,"total_runtime_s":61723.333333333336,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a19f3099038a34df3fa5","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2bd8a871e9d1ad9e5283c191","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23809523809523808,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":745,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[372];split=linux_only;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":745,"metadata":{"avgScore":0.4681714417182331,"avg_score":0.4681714417182331,"avg_score_percent":46.81714417182331,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":132.28,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":140230593.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":820328.5,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":50,"runs":205,"runtime":161026.859,"source_flags":[],"source_harness":"codex","source_row_index":372,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":132.28,"totalDurationS":161026.859,"totalInputTokens":140230593.5,"totalOutputTokens":820328.5,"total_cost_usd":132.28,"total_input_tokens":140230593.5,"total_output_tokens":820328.5,"total_runtime_s":161026.859,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.809523809523807,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-65f23e5451137689585a","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2a082d1e8d0a1cd1ea883219","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3930348258706468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1231,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[615];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1231,"metadata":{"avgScore":0.6677679137696316,"avg_score":0.6677679137696316,"avg_score_percent":66.77679137696316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.16,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":71044249.66666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":467625,"passRate":0.3930348258706468,"pass_rate":0.3930348258706468,"pass_rate_percent":39.30348258706468,"passes":53,"runs":134,"runtime":96179.83333333334,"source_flags":[],"source_harness":"codex","source_row_index":615,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":80.16,"totalDurationS":96179.83333333334,"totalInputTokens":71044249.66666667,"totalOutputTokens":467625,"total_cost_usd":80.16,"total_input_tokens":71044249.66666667,"total_output_tokens":467625,"total_runtime_s":96179.83333333334,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.30348258706468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0e9b949cea738435f1cd","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-277d0172a09ee252995cf0cb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.039473684210526314,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":247,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[123];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":247,"metadata":{"avgScore":0.19420078029613763,"avg_score":0.19420078029613763,"avg_score_percent":19.420078029613762,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":249.88,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":294882387,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":783893.5,"passRate":0.039473684210526314,"pass_rate":0.039473684210526314,"pass_rate_percent":3.9473684210526314,"passes":2,"runs":75,"runtime":117354.5,"source_flags":[],"source_harness":"codex","source_row_index":123,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":249.88,"totalDurationS":117354.5,"totalInputTokens":294882387,"totalOutputTokens":783893.5,"total_cost_usd":249.88,"total_input_tokens":294882387,"total_output_tokens":783893.5,"total_runtime_s":117354.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.9473684210526314,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-83675b153f37bac0afac","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2328a6da5ad59618ab8776dc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.05263157894736842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":251,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[125];split=full/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":251,"metadata":{"avgScore":0.1868668857038979,"avg_score":0.1868668857038979,"avg_score_percent":18.68668857038979,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":252.66,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":297522834,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1062195,"passRate":0.05263157894736842,"pass_rate":0.05263157894736842,"pass_rate_percent":5.263157894736842,"passes":3,"runs":75,"runtime":140665,"source_flags":[],"source_harness":"codex","source_row_index":125,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":252.66,"totalDurationS":140665,"totalInputTokens":297522834,"totalOutputTokens":1062195,"total_cost_usd":252.66,"total_input_tokens":297522834,"total_output_tokens":1062195,"total_runtime_s":140665,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":5.263157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b1983335e7f1fb5bd576","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-21faafb1bd8f95dfe090a718","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4552238805970149,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1193,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[596];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1193,"metadata":{"avgScore":0.7829241382323966,"avg_score":0.7829241382323966,"avg_score_percent":78.29241382323966,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":210.38,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":189978516.33333334,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1282200,"passRate":0.4552238805970149,"pass_rate":0.4552238805970149,"pass_rate_percent":45.52238805970149,"passes":61,"runs":134,"runtime":125006.5,"source_flags":[],"source_harness":"codex","source_row_index":596,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":210.38,"totalDurationS":125006.5,"totalInputTokens":189978516.33333334,"totalOutputTokens":1282200,"total_cost_usd":210.38,"total_input_tokens":189978516.33333334,"total_output_tokens":1282200,"total_runtime_s":125006.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.52238805970149,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bb1392fa55198b3df045","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2055270cf8ce69a8de51ce65","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2541371158392435,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1347,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[673];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1347,"metadata":{"avgScore":0.4681048909575152,"avg_score":0.4681048909575152,"avg_score_percent":46.81048909575152,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":223.43,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":230074698.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1175061.3333333333,"passRate":0.2541371158392435,"pass_rate":0.2541371158392435,"pass_rate_percent":25.413711583924346,"passes":72,"runs":278,"runtime":242446.52566666668,"source_flags":[],"source_harness":"codex","source_row_index":673,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":223.43,"totalDurationS":242446.52566666668,"totalInputTokens":230074698.16666666,"totalOutputTokens":1175061.3333333333,"total_cost_usd":223.43,"total_input_tokens":230074698.16666666,"total_output_tokens":1175061.3333333333,"total_runtime_s":242446.52566666668,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.413711583924346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0b4b45e8411b4f5b6a04","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1d4a7af64bba55210d242ffc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2390350877192982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":39,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[19];split=full;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":39,"metadata":{"avgScore":0.4509751802373446,"avg_score":0.4509751802373446,"avg_score_percent":45.09751802373446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":248.12,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":251962276.66666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":1269048.8333333333,"passRate":0.2390350877192982,"pass_rate":0.2390350877192982,"pass_rate_percent":23.903508771929822,"passes":73,"runs":299,"runtime":270652.0256666667,"source_flags":[],"source_harness":"codex","source_row_index":19,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":248.12,"totalDurationS":270652.0256666667,"totalInputTokens":251962276.66666666,"totalOutputTokens":1269048.8333333333,"total_cost_usd":248.12,"total_input_tokens":251962276.66666666,"total_output_tokens":1269048.8333333333,"total_runtime_s":270652.0256666667,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.903508771929822,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d4d2ad8f4afbcefbadb3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-194bf68633c7e599d33c5c58","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.317966903073286,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1317,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[658];split=unlicensed/overall;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1317,"metadata":{"avgScore":0.5430423343738866,"avg_score":0.5430423343738866,"avg_score_percent":54.30423343738866,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":424.24,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":447187737.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":2005900.5,"passRate":0.317966903073286,"pass_rate":0.317966903073286,"pass_rate_percent":31.7966903073286,"passes":87,"runs":280,"runtime":273534.44200000004,"source_flags":[],"source_harness":"codex","source_row_index":658,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":424.24,"totalDurationS":273534.44200000004,"totalInputTokens":447187737.1666667,"totalOutputTokens":2005900.5,"total_cost_usd":424.24,"total_input_tokens":447187737.1666667,"total_output_tokens":2005900.5,"total_runtime_s":273534.44200000004,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.7966903073286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6f7365c87e994c047cac","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-167aed8182f4d7cde1e74e2a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1115,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[557];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1115,"metadata":{"avgScore":0.12315914835731925,"avg_score":0.12315914835731925,"avg_score_percent":12.315914835731926,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":85.84,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":102937640,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":356578.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":58,"runtime":87469.83333333333,"source_flags":[],"source_harness":"codex","source_row_index":557,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":85.84,"totalDurationS":87469.83333333333,"totalInputTokens":102937640,"totalOutputTokens":356578.8333333333,"total_cost_usd":85.84,"total_input_tokens":102937640,"total_output_tokens":356578.8333333333,"total_runtime_s":87469.83333333333,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bf83cade9a7e5ccfdcf4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-137460004930b032f20610a1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.045454545454545456,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":611,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[305];split=licensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":611,"metadata":{"avgScore":0.23140343373334019,"avg_score":0.23140343373334019,"avg_score_percent":23.14034337333402,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":24.69,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":21887578.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":93987.5,"passRate":0.045454545454545456,"pass_rate":0.045454545454545456,"pass_rate_percent":4.545454545454546,"passes":1,"runs":21,"runtime":28205.5,"source_flags":[],"source_harness":"codex","source_row_index":305,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":24.69,"totalDurationS":28205.5,"totalInputTokens":21887578.5,"totalOutputTokens":93987.5,"total_cost_usd":24.69,"total_input_tokens":21887578.5,"total_output_tokens":93987.5,"total_runtime_s":28205.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":4.545454545454546,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-436b611a2aabc4d37384","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0732bf78e98e035e3f5fccfa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.45273631840796025,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":363,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[181];split=full/near-term;harness=codex;model=GPT-5.6-Sol;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":363,"metadata":{"avgScore":0.7561914737576828,"avg_score":0.7561914737576828,"avg_score_percent":75.61914737576828,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":136.75,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":124228619.66666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":787503.5,"passRate":0.45273631840796025,"pass_rate":0.45273631840796025,"pass_rate_percent":45.27363184079603,"passes":60,"runs":134,"runtime":123361.5,"source_flags":[],"source_harness":"codex","source_row_index":181,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":136.75,"totalDurationS":123361.5,"totalInputTokens":124228619.66666667,"totalOutputTokens":787503.5,"total_cost_usd":136.75,"total_input_tokens":124228619.66666667,"total_output_tokens":787503.5,"total_runtime_s":123361.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.27363184079603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4573b98abe957f2a5678","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0652dbe86dd774c342b68bbc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Sol","modelRef":"GPT-5.6-Sol","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3262411347517731,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11ac860bd468c4454ad8594e708a45055013a45d","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":837,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[418];split=unlicensed;harness=codex;model=GPT-5.6-Sol;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":837,"metadata":{"avgScore":0.5595483966613978,"avg_score":0.5595483966613978,"avg_score_percent":55.954839666139776,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":656.72,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":652364022.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Sol","output_tokens":3478380.3333333335,"passRate":0.3262411347517731,"pass_rate":0.3262411347517731,"pass_rate_percent":32.62411347517731,"passes":92,"runs":280,"runtime":291953.19649999996,"source_flags":[],"source_harness":"codex","source_row_index":418,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":656.72,"totalDurationS":291953.19649999996,"totalInputTokens":652364022.1666666,"totalOutputTokens":3478380.3333333335,"total_cost_usd":656.72,"total_input_tokens":652364022.1666666,"total_output_tokens":3478380.3333333335,"total_runtime_s":291953.19649999996,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.62411347517731,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dec8c2714365410a8b49","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fb3ad18a8cd2cae267ada7a9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.42499288919011063,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1372,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[685];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1372,"metadata":{"avgScore":0.42499288919011063,"avg_score":0.42499288919011063,"avg_score_percent":42.499288919011065,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.42,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":153532999,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1050240.5,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":60,"runs":278,"runtime":250208.747,"source_flags":[],"source_harness":"codex","source_row_index":685,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":79.42,"totalDurationS":250208.747,"totalInputTokens":153532999,"totalOutputTokens":1050240.5,"total_cost_usd":79.42,"total_input_tokens":153532999,"total_output_tokens":1050240.5,"total_runtime_s":250208.747,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.499288919011065,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e9d9079518b5ea4cc3d0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f5e4474e3c98d3f09a5f30d6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.48609115476187026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":868,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[433];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":868,"metadata":{"avgScore":0.48609115476187026,"avg_score":0.48609115476187026,"avg_score_percent":48.60911547618703,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":233.48,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":480691367,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2592990,"passRate":0.2801418439716312,"pass_rate":0.2801418439716312,"pass_rate_percent":28.01418439716312,"passes":77,"runs":276,"runtime":303490.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":433,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":233.48,"totalDurationS":303490.11950000003,"totalInputTokens":480691367,"totalOutputTokens":2592990,"total_cost_usd":233.48,"total_input_tokens":480691367,"total_output_tokens":2592990,"total_runtime_s":303490.11950000003,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.60911547618703,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-98be0ec1b9a71a67243e","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f10f22c67db092be174db667","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6595062733569761,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":398,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[198];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":398,"metadata":{"avgScore":0.6595062733569761,"avg_score":0.6595062733569761,"avg_score_percent":65.9506273356976,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":47.32,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":84622600.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":590880,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":52,"runs":131,"runtime":130482.5,"source_flags":[],"source_harness":"codex","source_row_index":198,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":47.32,"totalDurationS":130482.5,"totalInputTokens":84622600.5,"totalOutputTokens":590880,"total_cost_usd":47.32,"total_input_tokens":84622600.5,"total_output_tokens":590880,"total_runtime_s":130482.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":65.9506273356976,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-739b0e91d7070d039e80","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e3329eb381e9357428fa2c4c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7327376744045965,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1206,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[602];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1206,"metadata":{"avgScore":0.7327376744045965,"avg_score":0.7327376744045965,"avg_score_percent":73.27376744045965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":141.11,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":268565325.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1956342.5,"passRate":0.417910447761194,"pass_rate":0.417910447761194,"pass_rate_percent":41.7910447761194,"passes":54,"runs":132,"runtime":133954.5,"source_flags":[],"source_harness":"codex","source_row_index":602,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":141.11,"totalDurationS":133954.5,"totalInputTokens":268565325.5,"totalOutputTokens":1956342.5,"total_cost_usd":141.11,"total_input_tokens":268565325.5,"total_output_tokens":1956342.5,"total_runtime_s":133954.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":73.27376744045965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cf96c14157c65fccab19","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ddcb4b1adeff6e1175de4615","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2728171107655029,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":602,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[300];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":602,"metadata":{"avgScore":0.2728171107655029,"avg_score":0.2728171107655029,"avg_score_percent":27.28171107655029,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.94,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":96180035.66666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":392277.8333333333,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":20,"runtime":91366,"source_flags":[],"source_harness":"codex","source_row_index":300,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":44.94,"totalDurationS":91366,"totalInputTokens":96180035.66666667,"totalOutputTokens":392277.8333333333,"total_cost_usd":44.94,"total_input_tokens":96180035.66666667,"total_output_tokens":392277.8333333333,"total_runtime_s":91366,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.28171107655029,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c072ad675efb958e8a60","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-dd64555eeced27263a3dbeea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4024908249653849,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":66,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[32];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":66,"metadata":{"avgScore":0.4024908249653849,"avg_score":0.4024908249653849,"avg_score_percent":40.24908249653849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.49,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":171679764,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1124088.5,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":60,"runs":300,"runtime":275061.747,"source_flags":[],"source_harness":"codex","source_row_index":32,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":89.49,"totalDurationS":275061.747,"totalInputTokens":171679764,"totalOutputTokens":1124088.5,"total_cost_usd":89.49,"total_input_tokens":171679764,"total_output_tokens":1124088.5,"total_runtime_s":275061.747,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.24908249653849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1f9fa3c0fcfa732af7c6","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-dbce6538c4c4e95f19e6ecd1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4438320139384285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1356,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[677];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1356,"metadata":{"avgScore":0.4438320139384285,"avg_score":0.4438320139384285,"avg_score_percent":44.38320139384285,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":111.31,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":213612416.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1450871,"passRate":0.24468085106382978,"pass_rate":0.24468085106382978,"pass_rate_percent":24.46808510638298,"passes":69,"runs":276,"runtime":284946.718,"source_flags":[],"source_harness":"codex","source_row_index":677,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":111.31,"totalDurationS":284946.718,"totalInputTokens":213612416.5,"totalOutputTokens":1450871,"total_cost_usd":111.31,"total_input_tokens":213612416.5,"total_output_tokens":1450871,"total_runtime_s":284946.718,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.38320139384285,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-09f7d8d3b07c02c7144a","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d969092417fd9cc65dfcbc21","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5077329832154835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":860,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[429];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":860,"metadata":{"avgScore":0.5077329832154835,"avg_score":0.5077329832154835,"avg_score_percent":50.77329832154835,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":331.24,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":720897948,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":3526337,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":82,"runs":277,"runtime":260404,"source_flags":[],"source_harness":"codex","source_row_index":429,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":331.24,"totalDurationS":260404,"totalInputTokens":720897948,"totalOutputTokens":3526337,"total_cost_usd":331.24,"total_input_tokens":720897948,"total_output_tokens":3526337,"total_runtime_s":260404,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.77329832154835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4d7fdb3c321ea7bb070b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d3ef53f45e1a1b5699c043b4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4133817997254725,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":150,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[74];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":150,"metadata":{"avgScore":0.4133817997254725,"avg_score":0.4133817997254725,"avg_score_percent":41.33817997254725,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.55,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":215779432,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1228943,"passRate":0.24545454545454545,"pass_rate":0.24545454545454545,"pass_rate_percent":24.545454545454547,"passes":27,"runs":110,"runtime":104391.5,"source_flags":[],"source_harness":"codex","source_row_index":74,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":107.55,"totalDurationS":104391.5,"totalInputTokens":215779432,"totalOutputTokens":1228943,"total_cost_usd":107.55,"total_input_tokens":215779432,"total_output_tokens":1228943,"total_runtime_s":104391.5,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.33817997254725,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cb5d54a1f4f570601585","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d2f73d6b6ef726926dd6ab59","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.439100891470296,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":762,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[380];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":762,"metadata":{"avgScore":0.439100891470296,"avg_score":0.439100891470296,"avg_score_percent":43.9100891470296,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":64.73,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":125054108,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1031431,"passRate":0.22380952380952382,"pass_rate":0.22380952380952382,"pass_rate_percent":22.380952380952383,"passes":47,"runs":207,"runtime":191221.218,"source_flags":[],"source_harness":"codex","source_row_index":380,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":64.73,"totalDurationS":191221.218,"totalInputTokens":125054108,"totalOutputTokens":1031431,"total_cost_usd":64.73,"total_input_tokens":125054108,"total_output_tokens":1031431,"total_runtime_s":191221.218,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.9100891470296,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6db7abc101badc138346","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d14a4b4c36bdcdf025f475d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10630694231146019,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1124,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[561];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1124,"metadata":{"avgScore":0.10630694231146019,"avg_score":0.10630694231146019,"avg_score_percent":10.63069423114602,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":87.9,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":204857419.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":845060.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":56,"runtime":75359.5,"source_flags":[],"source_harness":"codex","source_row_index":561,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":87.9,"totalDurationS":75359.5,"totalInputTokens":204857419.5,"totalOutputTokens":845060.5,"total_cost_usd":87.9,"total_input_tokens":204857419.5,"total_output_tokens":845060.5,"total_runtime_s":75359.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.63069423114602,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7d3349f31bd7233c1a0e","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cf7669554cb5a18f9f3e7edc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.42524049731525787,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":770,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[384];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":770,"metadata":{"avgScore":0.42524049731525787,"avg_score":0.42524049731525787,"avg_score_percent":42.524049731525785,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":50.45,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":99257724.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":779045,"passRate":0.20476190476190476,"pass_rate":0.20476190476190476,"pass_rate_percent":20.476190476190474,"passes":43,"runs":208,"runtime":164807.247,"source_flags":[],"source_harness":"codex","source_row_index":384,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":50.45,"totalDurationS":164807.247,"totalInputTokens":99257724.5,"totalOutputTokens":779045,"total_cost_usd":50.45,"total_input_tokens":99257724.5,"total_output_tokens":779045,"total_runtime_s":164807.247,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.524049731525785,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-df81c8e6f5826ce570d0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c5c2b8c378b3fc06d0c4f58c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10511247258063579,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1126,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[562];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1126,"metadata":{"avgScore":0.10511247258063579,"avg_score":0.10511247258063579,"avg_score_percent":10.511247258063579,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":23.24,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":51763892.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":279262.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":59,"runtime":73434,"source_flags":[],"source_harness":"codex","source_row_index":562,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":23.24,"totalDurationS":73434,"totalInputTokens":51763892.5,"totalOutputTokens":279262.5,"total_cost_usd":23.24,"total_input_tokens":51763892.5,"total_output_tokens":279262.5,"total_runtime_s":73434,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.511247258063579,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1221fac1d2715d7c851","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c4494571e871635c675c4abe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.39558362254691226,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":988,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[493];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":988,"metadata":{"avgScore":0.39558362254691226,"avg_score":0.39558362254691226,"avg_score_percent":39.558362254691225,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":67.99,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":131879860,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":836993,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":21,"runs":99,"runtime":93679.6195,"source_flags":[],"source_harness":"codex","source_row_index":493,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":67.99,"totalDurationS":93679.6195,"totalInputTokens":131879860,"totalOutputTokens":836993,"total_cost_usd":67.99,"total_input_tokens":131879860,"total_output_tokens":836993,"total_runtime_s":93679.6195,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.558362254691225,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-252338c7a7283211af41","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c42a6dc4ae86352121a5d83d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5252918479026842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":854,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[426];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":854,"metadata":{"avgScore":0.5252918479026842,"avg_score":0.5252918479026842,"avg_score_percent":52.529184790268424,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":499.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1147950590,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":5390942,"passRate":0.29432624113475175,"pass_rate":0.29432624113475175,"pass_rate_percent":29.432624113475175,"passes":81,"runs":277,"runtime":335753,"source_flags":[],"source_harness":"codex","source_row_index":426,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":499.67,"totalDurationS":335753,"totalInputTokens":1147950590,"totalOutputTokens":5390942,"total_cost_usd":499.67,"total_input_tokens":1147950590,"total_output_tokens":5390942,"total_runtime_s":335753,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.529184790268424,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b87ef63f1edb1b5ea003","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c3c2dc7f0f2bee1dd76374eb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.11983054733814444,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":280,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[139];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":280,"metadata":{"avgScore":0.11983054733814444,"avg_score":0.11983054733814444,"avg_score_percent":11.983054733814443,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":105.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":236864478.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":967459,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":72,"runtime":102194.5,"source_flags":[],"source_harness":"codex","source_row_index":139,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":105.65,"totalDurationS":102194.5,"totalInputTokens":236864478.5,"totalOutputTokens":967459,"total_cost_usd":105.65,"total_input_tokens":236864478.5,"total_output_tokens":967459,"total_runtime_s":102194.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.983054733814443,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-207596f4e39946de6263","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c334e8d5e68ce6af99fe29f8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10286059924850519,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1132,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[565];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1132,"metadata":{"avgScore":0.10286059924850519,"avg_score":0.10286059924850519,"avg_score_percent":10.286059924850518,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":32.32,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":71099685.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":415960,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":57,"runtime":74813.5,"source_flags":[],"source_harness":"codex","source_row_index":565,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":32.32,"totalDurationS":74813.5,"totalInputTokens":71099685.5,"totalOutputTokens":415960,"total_cost_usd":32.32,"total_input_tokens":71099685.5,"total_output_tokens":415960,"total_runtime_s":74813.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.286059924850518,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5f0a70357b3b7a095032","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bc1f1f09f5e54fd87cf6f9c5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4512691704888627,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":962,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[480];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":962,"metadata":{"avgScore":0.4512691704888627,"avg_score":0.4512691704888627,"avg_score_percent":45.12691704888627,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":129.2,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":269660670.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1641035.5,"passRate":0.26,"pass_rate":0.26,"pass_rate_percent":26.0,"passes":26,"runs":99,"runtime":94791.5,"source_flags":[],"source_harness":"codex","source_row_index":480,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":129.2,"totalDurationS":94791.5,"totalInputTokens":269660670.5,"totalOutputTokens":1641035.5,"total_cost_usd":129.2,"total_input_tokens":269660670.5,"total_output_tokens":1641035.5,"total_runtime_s":94791.5,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.12691704888627,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1437e14675143244720","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b8f8817b1cb44623bd1b24ae","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7387174887795719,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":368,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[183];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":368,"metadata":{"avgScore":0.7387174887795719,"avg_score":0.7387174887795719,"avg_score_percent":73.87174887795719,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.88,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":170739006.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1313492.5,"passRate":0.44776119402985076,"pass_rate":0.44776119402985076,"pass_rate_percent":44.776119402985074,"passes":58,"runs":132,"runtime":113197,"source_flags":[],"source_harness":"codex","source_row_index":183,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.88,"totalDurationS":113197,"totalInputTokens":170739006.5,"totalOutputTokens":1313492.5,"total_cost_usd":95.88,"total_input_tokens":170739006.5,"total_output_tokens":1313492.5,"total_runtime_s":113197,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":73.87174887795719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b51a8d4d7e02e7fd1249","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b8281423da7be4d5c9ebdfd9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.484942264470657,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":500,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[249];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":500,"metadata":{"avgScore":0.484942264470657,"avg_score":0.484942264470657,"avg_score_percent":48.4942264470657,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":380.73,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":822499178,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":3817073,"passRate":0.27631578947368424,"pass_rate":0.27631578947368424,"pass_rate_percent":27.631578947368425,"passes":82,"runs":299,"runtime":314395.5,"source_flags":[],"source_harness":"codex","source_row_index":249,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":380.73,"totalDurationS":314395.5,"totalInputTokens":822499178,"totalOutputTokens":3817073,"total_cost_usd":380.73,"total_input_tokens":822499178,"total_output_tokens":3817073,"total_runtime_s":314395.5,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.4942264470657,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1289c701c667bdf8c213","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b056e6568c9b5c1adf36d4cf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1256614965929625,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":278,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[138];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":278,"metadata":{"avgScore":0.1256614965929625,"avg_score":0.1256614965929625,"avg_score_percent":12.566149659296249,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.29,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":90624808.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":486691,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":72,"runtime":92032.5,"source_flags":[],"source_harness":"codex","source_row_index":138,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":43.29,"totalDurationS":92032.5,"totalInputTokens":90624808.5,"totalOutputTokens":486691,"total_cost_usd":43.29,"total_input_tokens":90624808.5,"total_output_tokens":486691,"total_runtime_s":92032.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.566149659296249,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ee59df900238854a88f3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a986cddd51aa03b098ec1425","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5070206498203882,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":490,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[244];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":490,"metadata":{"avgScore":0.5070206498203882,"avg_score":0.5070206498203882,"avg_score_percent":50.70206498203882,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":544.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1244130625.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":5783219.833333333,"passRate":0.27960526315789475,"pass_rate":0.27960526315789475,"pass_rate_percent":27.960526315789476,"passes":82,"runs":297,"runtime":427119,"source_flags":[],"source_harness":"codex","source_row_index":244,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":544.61,"totalDurationS":427119,"totalInputTokens":1244130625.6666667,"totalOutputTokens":5783219.833333333,"total_cost_usd":544.61,"total_input_tokens":1244130625.6666667,"total_output_tokens":5783219.833333333,"total_runtime_s":427119,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.70206498203882,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f97233a0eb4c746c8b4b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a7eca400b795aad27d6bf954","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.14914850377407576,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":262,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[130];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":262,"metadata":{"avgScore":0.14914850377407576,"avg_score":0.14914850377407576,"avg_score_percent":14.914850377407577,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":264.49,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":678056488.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2240925.3333333335,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":2,"runs":74,"runtime":183287.5,"source_flags":[],"source_harness":"codex","source_row_index":130,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":264.49,"totalDurationS":183287.5,"totalInputTokens":678056488.1666666,"totalOutputTokens":2240925.3333333335,"total_cost_usd":264.49,"total_input_tokens":678056488.1666666,"total_output_tokens":2240925.3333333335,"total_runtime_s":183287.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.914850377407577,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7b7e9ce8a302e0002fec","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a4815a50df0a7628ad495a34","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4936421663563447,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":732,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[365];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":732,"metadata":{"avgScore":0.4936421663563447,"avg_score":0.4936421663563447,"avg_score_percent":49.36421663563447,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":234.69,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":528038236.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2578406,"passRate":0.2619047619047619,"pass_rate":0.2619047619047619,"pass_rate_percent":26.190476190476193,"passes":54,"runs":206,"runtime":176760,"source_flags":[],"source_harness":"codex","source_row_index":365,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":234.69,"totalDurationS":176760,"totalInputTokens":528038236.5,"totalOutputTokens":2578406,"total_cost_usd":234.69,"total_input_tokens":528038236.5,"total_output_tokens":2578406,"total_runtime_s":176760,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.36421663563447,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0977ce17745b835ad41f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-94b67fee73821b645a7817c5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4024908249653849,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":538,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[268];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":538,"metadata":{"avgScore":0.4024908249653849,"avg_score":0.4024908249653849,"avg_score_percent":40.24908249653849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.49,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":171679764,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1124088.5,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":60,"runs":300,"runtime":275061.747,"source_flags":[],"source_harness":"codex","source_row_index":268,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":89.49,"totalDurationS":275061.747,"totalInputTokens":171679764,"totalOutputTokens":1124088.5,"total_cost_usd":89.49,"total_input_tokens":171679764,"total_output_tokens":1124088.5,"total_runtime_s":275061.747,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.24908249653849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-36f0326661c5925c56ff","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-92715cae4adebbf89d40e101","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5012199103904947,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":730,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[364];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":730,"metadata":{"avgScore":0.5012199103904947,"avg_score":0.5012199103904947,"avg_score_percent":50.121991039049476,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":357.6,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":843825855.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":3883120,"passRate":0.2571428571428571,"pass_rate":0.2571428571428571,"pass_rate_percent":25.71428571428571,"passes":53,"runs":206,"runtime":211284.5,"source_flags":[],"source_harness":"codex","source_row_index":364,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":357.6,"totalDurationS":211284.5,"totalInputTokens":843825855.5,"totalOutputTokens":3883120,"total_cost_usd":357.6,"total_input_tokens":843825855.5,"total_output_tokens":3883120,"total_runtime_s":211284.5,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.121991039049476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-58d8bea5daeea2b5b54c","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8cd0ed5a0cd559077cf2c250","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3390348819796931,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1022,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[510];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1022,"metadata":{"avgScore":0.3390348819796931,"avg_score":0.3390348819796931,"avg_score_percent":33.90348819796931,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":34.63,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":62817795.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":496519.5,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":18,"runs":100,"runtime":88605.218,"source_flags":[],"source_harness":"codex","source_row_index":510,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":34.63,"totalDurationS":88605.218,"totalInputTokens":62817795.5,"totalOutputTokens":496519.5,"total_cost_usd":34.63,"total_input_tokens":62817795.5,"total_output_tokens":496519.5,"total_runtime_s":88605.218,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":33.90348819796931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d7256b7c1b6a0a7e87fa","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8a0599eaa6db3ab9a31da57d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.484942264470657,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":28,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[13];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":28,"metadata":{"avgScore":0.484942264470657,"avg_score":0.484942264470657,"avg_score_percent":48.4942264470657,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":380.73,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":822499178,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":3817073,"passRate":0.27631578947368424,"pass_rate":0.27631578947368424,"pass_rate_percent":27.631578947368425,"passes":82,"runs":299,"runtime":314395.5,"source_flags":[],"source_harness":"codex","source_row_index":13,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":380.73,"totalDurationS":314395.5,"totalInputTokens":822499178,"totalOutputTokens":3817073,"total_cost_usd":380.73,"total_input_tokens":822499178,"total_output_tokens":3817073,"total_runtime_s":314395.5,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.4942264470657,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6d022a28f5ffe918a7cc","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8824f3bfad6abf6b1f564773","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6068743663111191,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1254,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[626];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1254,"metadata":{"avgScore":0.6068743663111191,"avg_score":0.6068743663111191,"avg_score_percent":60.687436631111915,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":33.74,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":62201618.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":448088.5,"passRate":0.3208955223880597,"pass_rate":0.3208955223880597,"pass_rate_percent":32.08955223880597,"passes":41,"runs":131,"runtime":116658,"source_flags":[],"source_harness":"codex","source_row_index":626,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":33.74,"totalDurationS":116658,"totalInputTokens":62201618.5,"totalOutputTokens":448088.5,"total_cost_usd":33.74,"total_input_tokens":62201618.5,"total_output_tokens":448088.5,"total_runtime_s":116658,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.687436631111915,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-53e056010b11442f21aa","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-856b36ef06c62e1c597fe0e7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.46188106408078433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":126,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[62];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":126,"metadata":{"avgScore":0.46188106408078433,"avg_score":0.46188106408078433,"avg_score_percent":46.18810640807843,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":149.75,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":316865419,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1818837,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":29,"runs":107,"runtime":137684.5,"source_flags":[],"source_harness":"codex","source_row_index":62,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":149.75,"totalDurationS":137684.5,"totalInputTokens":316865419,"totalOutputTokens":1818837,"total_cost_usd":149.75,"total_input_tokens":316865419,"total_output_tokens":1818837,"total_runtime_s":137684.5,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.18810640807843,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-247ac4add991e8ec49c3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-83e28f92b026a25e08a651f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7084704112498615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":382,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[190];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":382,"metadata":{"avgScore":0.7084704112498615,"avg_score":0.7084704112498615,"avg_score_percent":70.84704112498615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":83.8,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155958887.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1013526,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":57,"runs":133,"runtime":144963,"source_flags":[],"source_harness":"codex","source_row_index":190,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":83.8,"totalDurationS":144963,"totalInputTokens":155958887.5,"totalOutputTokens":1013526,"total_cost_usd":83.8,"total_input_tokens":155958887.5,"total_output_tokens":1013526,"total_runtime_s":144963,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":70.84704112498615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3af1657c458d4699cc59","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7973fe1458bfebb83f01ab94","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7084704112498615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1218,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[608];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1218,"metadata":{"avgScore":0.7084704112498615,"avg_score":0.7084704112498615,"avg_score_percent":70.84704112498615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":83.8,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155958887.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1013526,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":57,"runs":133,"runtime":144963,"source_flags":[],"source_harness":"codex","source_row_index":608,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":83.8,"totalDurationS":144963,"totalInputTokens":155958887.5,"totalOutputTokens":1013526,"total_cost_usd":83.8,"total_input_tokens":155958887.5,"total_output_tokens":1013526,"total_runtime_s":144963,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":70.84704112498615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ba1b640bc9730d12ec9c","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7737c2f17e9becdbaa0454c9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4148733134041199,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":984,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[491];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":984,"metadata":{"avgScore":0.4148733134041199,"avg_score":0.4148733134041199,"avg_score_percent":41.48733134041199,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":91.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":177458632.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1109713,"passRate":0.25,"pass_rate":0.25,"pass_rate_percent":25.0,"passes":25,"runs":100,"runtime":73476,"source_flags":[],"source_harness":"codex","source_row_index":491,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":91.53,"totalDurationS":73476,"totalInputTokens":177458632.5,"totalOutputTokens":1109713,"total_cost_usd":91.53,"total_input_tokens":177458632.5,"total_output_tokens":1109713,"total_runtime_s":73476,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.48733134041199,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-54e4ae8127c5e6c204a7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-73e4a243871ac86438988e31","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1122775555549231,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":286,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[142];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":286,"metadata":{"avgScore":0.1122775555549231,"avg_score":0.1122775555549231,"avg_score_percent":11.22775555549231,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":65899280,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":335902,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":75,"runtime":86350,"source_flags":[],"source_harness":"codex","source_row_index":142,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":31.18,"totalDurationS":86350,"totalInputTokens":65899280,"totalOutputTokens":335902,"total_cost_usd":31.18,"total_input_tokens":65899280,"total_output_tokens":335902,"total_runtime_s":86350,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.22775555549231,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-89949f3271b096e0cbe1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-729db638c65db81bdd630db0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3891446401864527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":162,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[80];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":162,"metadata":{"avgScore":0.3891446401864527,"avg_score":0.3891446401864527,"avg_score_percent":38.91446401864527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":83.43,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":164102242.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":940673.5,"passRate":0.20909090909090908,"pass_rate":0.20909090909090908,"pass_rate_percent":20.909090909090907,"passes":22,"runs":109,"runtime":131295.6195,"source_flags":[],"source_harness":"codex","source_row_index":80,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":83.43,"totalDurationS":131295.6195,"totalInputTokens":164102242.5,"totalOutputTokens":940673.5,"total_cost_usd":83.43,"total_input_tokens":164102242.5,"total_output_tokens":940673.5,"total_runtime_s":131295.6195,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.91446401864527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b1909e5e7020293cdda1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-71e6c70b63dfeceb7cf2ea38","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.20454787385590675,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":628,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[313];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":628,"metadata":{"avgScore":0.20454787385590675,"avg_score":0.20454787385590675,"avg_score_percent":20.454787385590674,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":28.04,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":53183956.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":200989.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":51418,"source_flags":[],"source_harness":"codex","source_row_index":313,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":28.04,"totalDurationS":51418,"totalInputTokens":53183956.5,"totalOutputTokens":200989.5,"total_cost_usd":28.04,"total_input_tokens":53183956.5,"total_output_tokens":200989.5,"total_runtime_s":51418,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.454787385590674,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-15f3baf4cb51b0390511","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-655625e35603c38e247dcbbc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.48609115476187026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1340,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[669];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1340,"metadata":{"avgScore":0.48609115476187026,"avg_score":0.48609115476187026,"avg_score_percent":48.60911547618703,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":233.48,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":480691367,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2592990,"passRate":0.2801418439716312,"pass_rate":0.2801418439716312,"pass_rate_percent":28.01418439716312,"passes":77,"runs":276,"runtime":303490.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":669,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":233.48,"totalDurationS":303490.11950000003,"totalInputTokens":480691367,"totalOutputTokens":2592990,"total_cost_usd":233.48,"total_input_tokens":480691367,"total_output_tokens":2592990,"total_runtime_s":303490.11950000003,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.60911547618703,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92d7aa345a42d4615377","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-637fe3966adcb838dee2a795","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.465716312064728,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":504,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[251];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":504,"metadata":{"avgScore":0.465716312064728,"avg_score":0.465716312064728,"avg_score_percent":46.5716312064728,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":261.53,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":533875323.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2793979.5,"passRate":0.2598684210526316,"pass_rate":0.2598684210526316,"pass_rate_percent":25.986842105263158,"passes":77,"runs":298,"runtime":354908.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":251,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":261.53,"totalDurationS":354908.11950000003,"totalInputTokens":533875323.5,"totalOutputTokens":2793979.5,"total_cost_usd":261.53,"total_input_tokens":533875323.5,"total_output_tokens":2793979.5,"total_runtime_s":354908.11950000003,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.5716312064728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6c8cf8b82edd09d25886","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5cc41db365835d7e1920bbc2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3399830316257406,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":184,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[91];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":184,"metadata":{"avgScore":0.3399830316257406,"avg_score":0.3399830316257406,"avg_score_percent":33.99830316257406,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.6,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":78942311.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":555237.5,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":20,"runs":109,"runtime":101329.218,"source_flags":[],"source_harness":"codex","source_row_index":91,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":43.6,"totalDurationS":101329.218,"totalInputTokens":78942311.5,"totalOutputTokens":555237.5,"total_cost_usd":43.6,"total_input_tokens":78942311.5,"total_output_tokens":555237.5,"total_runtime_s":101329.218,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":33.99830316257406,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e2f166496ab75354e015","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5a3466cda3046aceff2d5ab2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.42499288919011063,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":900,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[449];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":900,"metadata":{"avgScore":0.42499288919011063,"avg_score":0.42499288919011063,"avg_score_percent":42.499288919011065,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.42,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":153532999,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1050240.5,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":60,"runs":278,"runtime":250208.747,"source_flags":[],"source_harness":"codex","source_row_index":449,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":79.42,"totalDurationS":250208.747,"totalInputTokens":153532999,"totalOutputTokens":1050240.5,"total_cost_usd":79.42,"total_input_tokens":153532999,"total_output_tokens":1050240.5,"total_runtime_s":250208.747,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.499288919011065,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1930c49d22d66f6c32b7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-57a08955edadd871c0ebc891","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4688643981010808,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":744,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[371];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":744,"metadata":{"avgScore":0.4688643981010808,"avg_score":0.4688643981010808,"avg_score_percent":46.88643981010808,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":144.1,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":297435267,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1832875,"passRate":0.24285714285714285,"pass_rate":0.24285714285714285,"pass_rate_percent":24.285714285714285,"passes":51,"runs":207,"runtime":167006.6195,"source_flags":[],"source_harness":"codex","source_row_index":371,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":144.1,"totalDurationS":167006.6195,"totalInputTokens":297435267,"totalOutputTokens":1832875,"total_cost_usd":144.1,"total_input_tokens":297435267,"total_output_tokens":1832875,"total_runtime_s":167006.6195,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.88643981010808,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-65959195f4cfb585b77f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4c84754372ae77614744be12","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5077329832154835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1332,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[665];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1332,"metadata":{"avgScore":0.5077329832154835,"avg_score":0.5077329832154835,"avg_score_percent":50.77329832154835,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":331.24,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":720897948,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":3526337,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":82,"runs":277,"runtime":260404,"source_flags":[],"source_harness":"codex","source_row_index":665,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":331.24,"totalDurationS":260404,"totalInputTokens":720897948,"totalOutputTokens":3526337,"total_cost_usd":331.24,"total_input_tokens":720897948,"total_output_tokens":3526337,"total_runtime_s":260404,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.77329832154835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-10d7b80f548bb0cc106f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3b95ce3ff8225f8a8b08817f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.15022183083314491,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1100,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[549];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1100,"metadata":{"avgScore":0.15022183083314491,"avg_score":0.15022183083314491,"avg_score_percent":15.022183083314491,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":239.29,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":627914170,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2004620,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":2,"runs":58,"runtime":115749.5,"source_flags":[],"source_harness":"codex","source_row_index":549,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":239.29,"totalDurationS":115749.5,"totalInputTokens":627914170,"totalOutputTokens":2004620,"total_cost_usd":239.29,"total_input_tokens":627914170,"total_output_tokens":2004620,"total_runtime_s":115749.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":15.022183083314491,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8945c70a474d51e3845","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-338c89b63dbdf38bcc0b010f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3447683087393445,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":180,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[89];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":180,"metadata":{"avgScore":0.3447683087393445,"avg_score":0.3447683087393445,"avg_score_percent":34.476830873934446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":30.84,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":54493605,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":398065.5,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":20,"runs":110,"runtime":82708.247,"source_flags":[],"source_harness":"codex","source_row_index":89,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":30.84,"totalDurationS":82708.247,"totalInputTokens":54493605,"totalOutputTokens":398065.5,"total_cost_usd":30.84,"total_input_tokens":54493605,"total_output_tokens":398065.5,"total_runtime_s":82708.247,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":34.476830873934446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f15e1c7dbeb1ea149d38","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-316586f2c58c9f5eecd43f1f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5070206498203882,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":18,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[8];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":18,"metadata":{"avgScore":0.5070206498203882,"avg_score":0.5070206498203882,"avg_score_percent":50.70206498203882,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":544.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1244130625.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":5783219.833333333,"passRate":0.27960526315789475,"pass_rate":0.27960526315789475,"pass_rate_percent":27.960526315789476,"passes":82,"runs":297,"runtime":427119,"source_flags":[],"source_harness":"codex","source_row_index":8,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":544.61,"totalDurationS":427119,"totalInputTokens":1244130625.6666667,"totalOutputTokens":5783219.833333333,"total_cost_usd":544.61,"total_input_tokens":1244130625.6666667,"total_output_tokens":5783219.833333333,"total_runtime_s":427119,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.70206498203882,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e1e9dd8ca80922cca52","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3061d3e66122660e084a7e92","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19554013940986384,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":634,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[316];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":634,"metadata":{"avgScore":0.19554013940986384,"avg_score":0.19554013940986384,"avg_score_percent":19.554013940986383,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":16.92,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":31029753,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":114260.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":21,"runtime":25065.5,"source_flags":[],"source_harness":"codex","source_row_index":316,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":16.92,"totalDurationS":25065.5,"totalInputTokens":31029753,"totalOutputTokens":114260.5,"total_cost_usd":16.92,"total_input_tokens":31029753,"total_output_tokens":114260.5,"total_runtime_s":25065.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.554013940986383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a9477606f7321ab59e36","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2f348743862ab0473cd45d3c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4258635230185982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":46,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[22];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":46,"metadata":{"avgScore":0.4258635230185982,"avg_score":0.4258635230185982,"avg_score_percent":42.58635230185982,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":128.23,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":244642169.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1565131.5,"passRate":0.22697368421052633,"pass_rate":0.22697368421052633,"pass_rate_percent":22.697368421052634,"passes":69,"runs":297,"runtime":310012.218,"source_flags":[],"source_harness":"codex","source_row_index":22,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":128.23,"totalDurationS":310012.218,"totalInputTokens":244642169.5,"totalOutputTokens":1565131.5,"total_cost_usd":128.23,"total_input_tokens":244642169.5,"total_output_tokens":1565131.5,"total_runtime_s":310012.218,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.58635230185982,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b59851dc6fde7cd31ca1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2c0d8620b1c48194574f03f7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.36165443790798124,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1014,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[506];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1014,"metadata":{"avgScore":0.36165443790798124,"avg_score":0.36165443790798124,"avg_score_percent":36.16544379079812,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":24.84,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":43552329,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":365133,"passRate":0.19,"pass_rate":0.19,"pass_rate_percent":19.0,"passes":19,"runs":100,"runtime":67713.747,"source_flags":[],"source_harness":"codex","source_row_index":506,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":24.84,"totalDurationS":67713.747,"totalInputTokens":43552329,"totalOutputTokens":365133,"total_cost_usd":24.84,"total_input_tokens":43552329,"total_output_tokens":365133,"total_runtime_s":67713.747,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.16544379079812,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-baf31822b57d30487c2d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2a135b10372e6e2207f027ba","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4438320139384285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":884,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[441];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":884,"metadata":{"avgScore":0.4438320139384285,"avg_score":0.4438320139384285,"avg_score_percent":44.38320139384285,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":111.31,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":213612416.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1450871,"passRate":0.24468085106382978,"pass_rate":0.24468085106382978,"pass_rate_percent":24.46808510638298,"passes":69,"runs":276,"runtime":284946.718,"source_flags":[],"source_harness":"codex","source_row_index":441,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":111.31,"totalDurationS":284946.718,"totalInputTokens":213612416.5,"totalOutputTokens":1450871,"total_cost_usd":111.31,"total_input_tokens":213612416.5,"total_output_tokens":1450871,"total_runtime_s":284946.718,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.38320139384285,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-71f50f6fbed0c9035c8c","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1dd9d46a2d07642cd5fa6917","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.465716312064728,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":32,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[15];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":32,"metadata":{"avgScore":0.465716312064728,"avg_score":0.465716312064728,"avg_score_percent":46.5716312064728,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":261.53,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":533875323.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":2793979.5,"passRate":0.2598684210526316,"pass_rate":0.2598684210526316,"pass_rate_percent":25.986842105263158,"passes":77,"runs":298,"runtime":354908.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":15,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":261.53,"totalDurationS":354908.11950000003,"totalInputTokens":533875323.5,"totalOutputTokens":2793979.5,"total_cost_usd":261.53,"total_input_tokens":533875323.5,"total_output_tokens":2793979.5,"total_runtime_s":354908.11950000003,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.5716312064728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2b6d2d3a87a61d08ffe0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1b9113fff5be28b105084db3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6595062733569761,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1234,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[616];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1234,"metadata":{"avgScore":0.6595062733569761,"avg_score":0.6595062733569761,"avg_score_percent":65.9506273356976,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":47.32,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":84622600.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":590880,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":52,"runs":131,"runtime":130482.5,"source_flags":[],"source_harness":"codex","source_row_index":616,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":47.32,"totalDurationS":130482.5,"totalInputTokens":84622600.5,"totalOutputTokens":590880,"total_cost_usd":47.32,"total_input_tokens":84622600.5,"total_output_tokens":590880,"total_runtime_s":130482.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":65.9506273356976,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b51a2d2f7bdebf829d15","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-172ef306bf8cf457eedd2b7e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4258635230185982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":518,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[258];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":518,"metadata":{"avgScore":0.4258635230185982,"avg_score":0.4258635230185982,"avg_score_percent":42.58635230185982,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":128.23,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":244642169.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1565131.5,"passRate":0.22697368421052633,"pass_rate":0.22697368421052633,"pass_rate_percent":22.697368421052634,"passes":69,"runs":297,"runtime":310012.218,"source_flags":[],"source_harness":"codex","source_row_index":258,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":128.23,"totalDurationS":310012.218,"totalInputTokens":244642169.5,"totalOutputTokens":1565131.5,"total_cost_usd":128.23,"total_input_tokens":244642169.5,"total_output_tokens":1565131.5,"total_runtime_s":310012.218,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.58635230185982,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6ba1c06b9ca217ddb119","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-16d656cc4831906ea4ef291c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1092923011990062,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":292,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[145];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":292,"metadata":{"avgScore":0.1092923011990062,"avg_score":0.1092923011990062,"avg_score_percent":10.92923011990062,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":185.5,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":449191663,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1428521,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":73,"runtime":120135.5,"source_flags":[],"source_harness":"codex","source_row_index":145,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":185.5,"totalDurationS":120135.5,"totalInputTokens":449191663,"totalOutputTokens":1428521,"total_cost_usd":185.5,"total_input_tokens":449191663,"total_output_tokens":1428521,"total_runtime_s":120135.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.92923011990062,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f9e07155282b9c536fd0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-168be6fe2c9738dd4190668f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6068743663111191,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":418,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[208];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":418,"metadata":{"avgScore":0.6068743663111191,"avg_score":0.6068743663111191,"avg_score_percent":60.687436631111915,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":33.74,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":62201618.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":448088.5,"passRate":0.3208955223880597,"pass_rate":0.3208955223880597,"pass_rate_percent":32.08955223880597,"passes":41,"runs":131,"runtime":116658,"source_flags":[],"source_harness":"codex","source_row_index":208,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":33.74,"totalDurationS":116658,"totalInputTokens":62201618.5,"totalOutputTokens":448088.5,"total_cost_usd":33.74,"total_input_tokens":62201618.5,"total_output_tokens":448088.5,"total_runtime_s":116658,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.687436631111915,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c02e24863885d5911a94","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0ed4b513b44c6e98cc02ad13","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19280668783242483,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":636,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[317];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":636,"metadata":{"avgScore":0.19280668783242483,"avg_score":0.19280668783242483,"avg_score_percent":19.280668783242483,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.49,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":101601230,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":290736,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":53991.5,"source_flags":[],"source_harness":"codex","source_row_index":317,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":49.49,"totalDurationS":53991.5,"totalInputTokens":101601230,"totalOutputTokens":290736,"total_cost_usd":49.49,"total_input_tokens":101601230,"total_output_tokens":290736,"total_runtime_s":53991.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.280668783242483,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b4e5cfbee0b8259ce563","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0dbd495c0e2682bbf7892d79","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7327376744045965,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":370,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[184];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":370,"metadata":{"avgScore":0.7327376744045965,"avg_score":0.7327376744045965,"avg_score_percent":73.27376744045965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":141.11,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":268565325.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1956342.5,"passRate":0.417910447761194,"pass_rate":0.417910447761194,"pass_rate_percent":41.7910447761194,"passes":54,"runs":132,"runtime":133954.5,"source_flags":[],"source_harness":"codex","source_row_index":184,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":141.11,"totalDurationS":133954.5,"totalInputTokens":268565325.5,"totalOutputTokens":1956342.5,"total_cost_usd":141.11,"total_input_tokens":268565325.5,"total_output_tokens":1956342.5,"total_runtime_s":133954.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":73.27376744045965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a10afcb24c3ec5d4c9bc","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0c0f54a6e837c062edbdcb17","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.11405527444844456,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":662,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[330];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":662,"metadata":{"avgScore":0.11405527444844456,"avg_score":0.11405527444844456,"avg_score_percent":11.405527444844456,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":10.07,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":18146765,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":73848,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":24853,"source_flags":[],"source_harness":"codex","source_row_index":330,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":10.07,"totalDurationS":24853,"totalInputTokens":18146765,"totalOutputTokens":73848,"total_cost_usd":10.07,"total_input_tokens":18146765,"total_output_tokens":73848,"total_runtime_s":24853,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.405527444844456,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-907766b8eb6295890329","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0ae28914a40ec444d4eb19ca","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5252918479026842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1326,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[662];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1326,"metadata":{"avgScore":0.5252918479026842,"avg_score":0.5252918479026842,"avg_score_percent":52.529184790268424,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":499.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1147950590,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":5390942,"passRate":0.29432624113475175,"pass_rate":0.29432624113475175,"pass_rate_percent":29.432624113475175,"passes":81,"runs":277,"runtime":335753,"source_flags":[],"source_harness":"codex","source_row_index":662,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":499.67,"totalDurationS":335753,"totalInputTokens":1147950590,"totalOutputTokens":5390942,"total_cost_usd":499.67,"total_input_tokens":1147950590,"total_output_tokens":5390942,"total_runtime_s":335753,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.529184790268424,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5e738a39509509cdf4d7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0803b6400e46138d710b9660","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10081890647001854,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1134,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[566];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1134,"metadata":{"avgScore":0.10081890647001854,"avg_score":0.10081890647001854,"avg_score_percent":10.081890647001854,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":151.35,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":384973196,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1243671,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":57,"runtime":81127.5,"source_flags":[],"source_harness":"codex","source_row_index":566,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":151.35,"totalDurationS":81127.5,"totalInputTokens":384973196,"totalOutputTokens":1243671,"total_cost_usd":151.35,"total_input_tokens":384973196,"total_output_tokens":1243671,"total_runtime_s":81127.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.081890647001854,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c08ee2983c87417ba336","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-03b691c50737b10c840f985d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7387174887795719,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c6e192dde0e993991cb391824e16d5516a67746","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1204,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[601];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1204,"metadata":{"avgScore":0.7387174887795719,"avg_score":0.7387174887795719,"avg_score_percent":73.87174887795719,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.88,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":170739006.5,"metric_id":"avg_score","metric_role":"secondary","model":"GPT-5.6-Terra","output_tokens":1313492.5,"passRate":0.44776119402985076,"pass_rate":0.44776119402985076,"pass_rate_percent":44.776119402985074,"passes":58,"runs":132,"runtime":113197,"source_flags":[],"source_harness":"codex","source_row_index":601,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.88,"totalDurationS":113197,"totalInputTokens":170739006.5,"totalOutputTokens":1313492.5,"total_cost_usd":95.88,"total_input_tokens":170739006.5,"total_output_tokens":1313492.5,"total_runtime_s":113197,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":73.87174887795719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cc5a16c28e0a05338772","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fd21f1bd049ab64c5b0bfdc0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2801418439716312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1339,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[669];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1339,"metadata":{"avgScore":0.48609115476187026,"avg_score":0.48609115476187026,"avg_score_percent":48.60911547618703,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":233.48,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":480691367,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2592990,"passRate":0.2801418439716312,"pass_rate":0.2801418439716312,"pass_rate_percent":28.01418439716312,"passes":77,"runs":276,"runtime":303490.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":669,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":233.48,"totalDurationS":303490.11950000003,"totalInputTokens":480691367,"totalOutputTokens":2592990,"total_cost_usd":233.48,"total_input_tokens":480691367,"total_output_tokens":2592990,"total_runtime_s":303490.11950000003,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.01418439716312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c52d936ff2a0b3b1de96","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fc5f03c01a67d66cc33ad3fa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":183,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[91];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":183,"metadata":{"avgScore":0.3399830316257406,"avg_score":0.3399830316257406,"avg_score_percent":33.99830316257406,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.6,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":78942311.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":555237.5,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":20,"runs":109,"runtime":101329.218,"source_flags":[],"source_harness":"codex","source_row_index":91,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":43.6,"totalDurationS":101329.218,"totalInputTokens":78942311.5,"totalOutputTokens":555237.5,"total_cost_usd":43.6,"total_input_tokens":78942311.5,"total_output_tokens":555237.5,"total_runtime_s":101329.218,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d5f96f78ec3b15b93395","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f834b4910101be362da1c29e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44776119402985076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":367,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[183];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":367,"metadata":{"avgScore":0.7387174887795719,"avg_score":0.7387174887795719,"avg_score_percent":73.87174887795719,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.88,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":170739006.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1313492.5,"passRate":0.44776119402985076,"pass_rate":0.44776119402985076,"pass_rate_percent":44.776119402985074,"passes":58,"runs":132,"runtime":113197,"source_flags":[],"source_harness":"codex","source_row_index":183,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.88,"totalDurationS":113197,"totalInputTokens":170739006.5,"totalOutputTokens":1313492.5,"total_cost_usd":95.88,"total_input_tokens":170739006.5,"total_output_tokens":1313492.5,"total_runtime_s":113197,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.776119402985074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1f8876d7e3d31dc6306f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f2158361cfd73fcfa6ca1a07","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1021,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[510];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1021,"metadata":{"avgScore":0.3390348819796931,"avg_score":0.3390348819796931,"avg_score_percent":33.90348819796931,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":34.63,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":62817795.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":496519.5,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":18,"runs":100,"runtime":88605.218,"source_flags":[],"source_harness":"codex","source_row_index":510,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":34.63,"totalDurationS":88605.218,"totalInputTokens":62817795.5,"totalOutputTokens":496519.5,"total_cost_usd":34.63,"total_input_tokens":62817795.5,"total_output_tokens":496519.5,"total_runtime_s":88605.218,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-416743a74fcb72a16589","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f0561a2c4cc03cd7c1e67c49","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":179,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[89];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":179,"metadata":{"avgScore":0.3447683087393445,"avg_score":0.3447683087393445,"avg_score_percent":34.476830873934446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":30.84,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":54493605,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":398065.5,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":20,"runs":110,"runtime":82708.247,"source_flags":[],"source_harness":"codex","source_row_index":89,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":30.84,"totalDurationS":82708.247,"totalInputTokens":54493605,"totalOutputTokens":398065.5,"total_cost_usd":30.84,"total_input_tokens":54493605,"total_output_tokens":398065.5,"total_runtime_s":82708.247,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-092bc57ddddcc47ece83","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ef81dfed78def04332ffd6fe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.20394736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":537,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[268];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":537,"metadata":{"avgScore":0.4024908249653849,"avg_score":0.4024908249653849,"avg_score_percent":40.24908249653849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.49,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":171679764,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1124088.5,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":60,"runs":300,"runtime":275061.747,"source_flags":[],"source_harness":"codex","source_row_index":268,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":89.49,"totalDurationS":275061.747,"totalInputTokens":171679764,"totalOutputTokens":1124088.5,"total_cost_usd":89.49,"total_input_tokens":171679764,"total_output_tokens":1124088.5,"total_runtime_s":275061.747,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.394736842105264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1dbc9d5f1596ff86ec0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-eca9803be00acfe56d085b43","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2727272727272727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":125,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[62];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":125,"metadata":{"avgScore":0.46188106408078433,"avg_score":0.46188106408078433,"avg_score_percent":46.18810640807843,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":149.75,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":316865419,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1818837,"passRate":0.2727272727272727,"pass_rate":0.2727272727272727,"pass_rate_percent":27.27272727272727,"passes":29,"runs":107,"runtime":137684.5,"source_flags":[],"source_harness":"codex","source_row_index":62,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":149.75,"totalDurationS":137684.5,"totalInputTokens":316865419,"totalOutputTokens":1818837,"total_cost_usd":149.75,"total_input_tokens":316865419,"total_output_tokens":1818837,"total_runtime_s":137684.5,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6aa6c4de6084540f96d7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e881f05190d1406d80e36085","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.417910447761194,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":369,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[184];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":369,"metadata":{"avgScore":0.7327376744045965,"avg_score":0.7327376744045965,"avg_score_percent":73.27376744045965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":141.11,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":268565325.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1956342.5,"passRate":0.417910447761194,"pass_rate":0.417910447761194,"pass_rate_percent":41.7910447761194,"passes":54,"runs":132,"runtime":133954.5,"source_flags":[],"source_harness":"codex","source_row_index":184,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":141.11,"totalDurationS":133954.5,"totalInputTokens":268565325.5,"totalOutputTokens":1956342.5,"total_cost_usd":141.11,"total_input_tokens":268565325.5,"total_output_tokens":1956342.5,"total_runtime_s":133954.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.7910447761194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d305c186bf2eca931a20","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e68f959d888c204e8177fbe7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":277,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[138];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":277,"metadata":{"avgScore":0.1256614965929625,"avg_score":0.1256614965929625,"avg_score_percent":12.566149659296249,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.29,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":90624808.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":486691,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":72,"runtime":92032.5,"source_flags":[],"source_harness":"codex","source_row_index":138,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":43.29,"totalDurationS":92032.5,"totalInputTokens":90624808.5,"totalOutputTokens":486691,"total_cost_usd":43.29,"total_input_tokens":90624808.5,"total_output_tokens":486691,"total_runtime_s":92032.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-025b56b24516244fe458","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e686587fd60fe5d903466558","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2571428571428571,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":729,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[364];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":729,"metadata":{"avgScore":0.5012199103904947,"avg_score":0.5012199103904947,"avg_score_percent":50.121991039049476,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":357.6,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":843825855.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":3883120,"passRate":0.2571428571428571,"pass_rate":0.2571428571428571,"pass_rate_percent":25.71428571428571,"passes":53,"runs":206,"runtime":211284.5,"source_flags":[],"source_harness":"codex","source_row_index":364,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":357.6,"totalDurationS":211284.5,"totalInputTokens":843825855.5,"totalOutputTokens":3883120,"total_cost_usd":357.6,"total_input_tokens":843825855.5,"total_output_tokens":3883120,"total_runtime_s":211284.5,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.71428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8df13d9b99f1ac57d4f2","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e3404318923664885f967758","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44776119402985076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1203,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[601];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1203,"metadata":{"avgScore":0.7387174887795719,"avg_score":0.7387174887795719,"avg_score_percent":73.87174887795719,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":95.88,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":170739006.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1313492.5,"passRate":0.44776119402985076,"pass_rate":0.44776119402985076,"pass_rate_percent":44.776119402985074,"passes":58,"runs":132,"runtime":113197,"source_flags":[],"source_harness":"codex","source_row_index":601,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":95.88,"totalDurationS":113197,"totalInputTokens":170739006.5,"totalOutputTokens":1313492.5,"total_cost_usd":95.88,"total_input_tokens":170739006.5,"total_output_tokens":1313492.5,"total_runtime_s":113197,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.776119402985074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-873cec99e1cf0a1ce738","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e218f8d48192219d793e231e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.20476190476190476,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":769,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[384];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":769,"metadata":{"avgScore":0.42524049731525787,"avg_score":0.42524049731525787,"avg_score_percent":42.524049731525785,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":50.45,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":99257724.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":779045,"passRate":0.20476190476190476,"pass_rate":0.20476190476190476,"pass_rate_percent":20.476190476190474,"passes":43,"runs":208,"runtime":164807.247,"source_flags":[],"source_harness":"codex","source_row_index":384,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":50.45,"totalDurationS":164807.247,"totalInputTokens":99257724.5,"totalOutputTokens":779045,"total_cost_usd":50.45,"total_input_tokens":99257724.5,"total_output_tokens":779045,"total_runtime_s":164807.247,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.476190476190474,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-648fdd4e3cae9ad08d41","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-df7fb2062845a2f5b50f4ab6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.27960526315789475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":489,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[244];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":489,"metadata":{"avgScore":0.5070206498203882,"avg_score":0.5070206498203882,"avg_score_percent":50.70206498203882,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":544.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1244130625.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":5783219.833333333,"passRate":0.27960526315789475,"pass_rate":0.27960526315789475,"pass_rate_percent":27.960526315789476,"passes":82,"runs":297,"runtime":427119,"source_flags":[],"source_harness":"codex","source_row_index":244,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":544.61,"totalDurationS":427119,"totalInputTokens":1244130625.6666667,"totalOutputTokens":5783219.833333333,"total_cost_usd":544.61,"total_input_tokens":1244130625.6666667,"total_output_tokens":5783219.833333333,"total_runtime_s":427119,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.960526315789476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b83328c98ad6a218f2b1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-df6bb4bc1d5b657cb0878491","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1233,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[616];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1233,"metadata":{"avgScore":0.6595062733569761,"avg_score":0.6595062733569761,"avg_score_percent":65.9506273356976,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":47.32,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":84622600.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":590880,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":52,"runs":131,"runtime":130482.5,"source_flags":[],"source_harness":"codex","source_row_index":616,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":47.32,"totalDurationS":130482.5,"totalInputTokens":84622600.5,"totalOutputTokens":590880,"total_cost_usd":47.32,"total_input_tokens":84622600.5,"total_output_tokens":590880,"total_runtime_s":130482.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-15957e6f343d5d0b7540","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-da9c3642f00c5849962ff2c4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3208955223880597,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1253,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[626];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1253,"metadata":{"avgScore":0.6068743663111191,"avg_score":0.6068743663111191,"avg_score_percent":60.687436631111915,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":33.74,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":62201618.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":448088.5,"passRate":0.3208955223880597,"pass_rate":0.3208955223880597,"pass_rate_percent":32.08955223880597,"passes":41,"runs":131,"runtime":116658,"source_flags":[],"source_harness":"codex","source_row_index":626,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":33.74,"totalDurationS":116658,"totalInputTokens":62201618.5,"totalOutputTokens":448088.5,"total_cost_usd":33.74,"total_input_tokens":62201618.5,"total_output_tokens":448088.5,"total_runtime_s":116658,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.08955223880597,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-56a3ed47fbd42cf1b128","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-da311444a1dffe9859fbfff6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.417910447761194,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1205,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[602];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1205,"metadata":{"avgScore":0.7327376744045965,"avg_score":0.7327376744045965,"avg_score_percent":73.27376744045965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":141.11,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":268565325.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1956342.5,"passRate":0.417910447761194,"pass_rate":0.417910447761194,"pass_rate_percent":41.7910447761194,"passes":54,"runs":132,"runtime":133954.5,"source_flags":[],"source_harness":"codex","source_row_index":602,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":141.11,"totalDurationS":133954.5,"totalInputTokens":268565325.5,"totalOutputTokens":1956342.5,"total_cost_usd":141.11,"total_input_tokens":268565325.5,"total_output_tokens":1956342.5,"total_runtime_s":133954.5,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.7910447761194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e38b3c1fb6f6b1c9e6a9","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d397dde067c3660e51107095","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":601,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[300];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":601,"metadata":{"avgScore":0.2728171107655029,"avg_score":0.2728171107655029,"avg_score_percent":27.28171107655029,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.94,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":96180035.66666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":392277.8333333333,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":20,"runtime":91366,"source_flags":[],"source_harness":"codex","source_row_index":300,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":44.94,"totalDurationS":91366,"totalInputTokens":96180035.66666667,"totalOutputTokens":392277.8333333333,"total_cost_usd":44.94,"total_input_tokens":96180035.66666667,"total_output_tokens":392277.8333333333,"total_runtime_s":91366,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-747baa23544d69213ee0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d2382428c1378b5b4d9c2b6d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.20909090909090908,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":161,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[80];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":161,"metadata":{"avgScore":0.3891446401864527,"avg_score":0.3891446401864527,"avg_score_percent":38.91446401864527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":83.43,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":164102242.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":940673.5,"passRate":0.20909090909090908,"pass_rate":0.20909090909090908,"pass_rate_percent":20.909090909090907,"passes":22,"runs":109,"runtime":131295.6195,"source_flags":[],"source_harness":"codex","source_row_index":80,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":83.43,"totalDurationS":131295.6195,"totalInputTokens":164102242.5,"totalOutputTokens":940673.5,"total_cost_usd":83.43,"total_input_tokens":164102242.5,"total_output_tokens":940673.5,"total_runtime_s":131295.6195,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.909090909090907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-514bc266be22fa747c56","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d0c71eb437a0f832704538ea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1133,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[566];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1133,"metadata":{"avgScore":0.10081890647001854,"avg_score":0.10081890647001854,"avg_score_percent":10.081890647001854,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":151.35,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":384973196,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1243671,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":57,"runtime":81127.5,"source_flags":[],"source_harness":"codex","source_row_index":566,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":151.35,"totalDurationS":81127.5,"totalInputTokens":384973196,"totalOutputTokens":1243671,"total_cost_usd":151.35,"total_input_tokens":384973196,"total_output_tokens":1243671,"total_runtime_s":81127.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6047a742b620fcf1b0f6","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-be14d6e848b7743cea36771a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24545454545454545,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":149,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[74];split=full/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":149,"metadata":{"avgScore":0.4133817997254725,"avg_score":0.4133817997254725,"avg_score_percent":41.33817997254725,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.55,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":215779432,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1228943,"passRate":0.24545454545454545,"pass_rate":0.24545454545454545,"pass_rate_percent":24.545454545454547,"passes":27,"runs":110,"runtime":104391.5,"source_flags":[],"source_harness":"codex","source_row_index":74,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":107.55,"totalDurationS":104391.5,"totalInputTokens":215779432,"totalOutputTokens":1228943,"total_cost_usd":107.55,"total_input_tokens":215779432,"total_output_tokens":1228943,"total_runtime_s":104391.5,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.545454545454547,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-964315d0d0caacc4fe04","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ba6cdaa66f13162da4b0f52f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22697368421052633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":45,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[22];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":45,"metadata":{"avgScore":0.4258635230185982,"avg_score":0.4258635230185982,"avg_score_percent":42.58635230185982,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":128.23,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":244642169.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1565131.5,"passRate":0.22697368421052633,"pass_rate":0.22697368421052633,"pass_rate_percent":22.697368421052634,"passes":69,"runs":297,"runtime":310012.218,"source_flags":[],"source_harness":"codex","source_row_index":22,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":128.23,"totalDurationS":310012.218,"totalInputTokens":244642169.5,"totalOutputTokens":1565131.5,"total_cost_usd":128.23,"total_input_tokens":244642169.5,"total_output_tokens":1565131.5,"total_runtime_s":310012.218,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.697368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aec0e5cb92af63c693a5","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ad471688defd5235b43d488b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.27631578947368424,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":27,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[13];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":27,"metadata":{"avgScore":0.484942264470657,"avg_score":0.484942264470657,"avg_score_percent":48.4942264470657,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":380.73,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":822499178,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":3817073,"passRate":0.27631578947368424,"pass_rate":0.27631578947368424,"pass_rate_percent":27.631578947368425,"passes":82,"runs":299,"runtime":314395.5,"source_flags":[],"source_harness":"codex","source_row_index":13,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":380.73,"totalDurationS":314395.5,"totalInputTokens":822499178,"totalOutputTokens":3817073,"total_cost_usd":380.73,"total_input_tokens":822499178,"total_output_tokens":3817073,"total_runtime_s":314395.5,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.631578947368425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-df5d2237964c662ab4a9","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a99e505a878d6d5109a1b44a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.27960526315789475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":17,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[8];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":17,"metadata":{"avgScore":0.5070206498203882,"avg_score":0.5070206498203882,"avg_score_percent":50.70206498203882,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":544.61,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1244130625.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":5783219.833333333,"passRate":0.27960526315789475,"pass_rate":0.27960526315789475,"pass_rate_percent":27.960526315789476,"passes":82,"runs":297,"runtime":427119,"source_flags":[],"source_harness":"codex","source_row_index":8,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":544.61,"totalDurationS":427119,"totalInputTokens":1244130625.6666667,"totalOutputTokens":5783219.833333333,"total_cost_usd":544.61,"total_input_tokens":1244130625.6666667,"total_output_tokens":5783219.833333333,"total_runtime_s":427119,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.960526315789476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4720011f7cac59dfa595","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a6b8556ee6cdaf5e12c3fa31","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2978723404255319,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1331,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[665];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1331,"metadata":{"avgScore":0.5077329832154835,"avg_score":0.5077329832154835,"avg_score_percent":50.77329832154835,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":331.24,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":720897948,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":3526337,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":82,"runs":277,"runtime":260404,"source_flags":[],"source_harness":"codex","source_row_index":665,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":331.24,"totalDurationS":260404,"totalInputTokens":720897948,"totalOutputTokens":3526337,"total_cost_usd":331.24,"total_input_tokens":720897948,"total_output_tokens":3526337,"total_runtime_s":260404,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.78723404255319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a2be174edd320d242954","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a097227f85b85976970aefed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2978723404255319,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":859,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[429];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":859,"metadata":{"avgScore":0.5077329832154835,"avg_score":0.5077329832154835,"avg_score_percent":50.77329832154835,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":331.24,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":720897948,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":3526337,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":82,"runs":277,"runtime":260404,"source_flags":[],"source_harness":"codex","source_row_index":429,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":331.24,"totalDurationS":260404,"totalInputTokens":720897948,"totalOutputTokens":3526337,"total_cost_usd":331.24,"total_input_tokens":720897948,"total_output_tokens":3526337,"total_runtime_s":260404,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.78723404255319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2eeb6b5f411dbd1e2d86","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a08870fd2a3bf891ddb3a1a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.19,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1013,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[506];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1013,"metadata":{"avgScore":0.36165443790798124,"avg_score":0.36165443790798124,"avg_score_percent":36.16544379079812,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":24.84,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":43552329,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":365133,"passRate":0.19,"pass_rate":0.19,"pass_rate_percent":19.0,"passes":19,"runs":100,"runtime":67713.747,"source_flags":[],"source_harness":"codex","source_row_index":506,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":24.84,"totalDurationS":67713.747,"totalInputTokens":43552329,"totalOutputTokens":365133,"total_cost_usd":24.84,"total_input_tokens":43552329,"total_output_tokens":365133,"total_runtime_s":67713.747,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-db329261928f65725c02","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-89d8be89b1da221b6a011819","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1099,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[549];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1099,"metadata":{"avgScore":0.15022183083314491,"avg_score":0.15022183083314491,"avg_score_percent":15.022183083314491,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":239.29,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":627914170,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2004620,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":2,"runs":58,"runtime":115749.5,"source_flags":[],"source_harness":"codex","source_row_index":549,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":239.29,"totalDurationS":115749.5,"totalInputTokens":627914170,"totalOutputTokens":2004620,"total_cost_usd":239.29,"total_input_tokens":627914170,"total_output_tokens":2004620,"total_runtime_s":115749.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-133e42f9f5fa87d84ccb","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-88fd6d7e50bc2e2691ebac74","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22380952380952382,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":761,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[380];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":761,"metadata":{"avgScore":0.439100891470296,"avg_score":0.439100891470296,"avg_score_percent":43.9100891470296,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":64.73,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":125054108,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1031431,"passRate":0.22380952380952382,"pass_rate":0.22380952380952382,"pass_rate_percent":22.380952380952383,"passes":47,"runs":207,"runtime":191221.218,"source_flags":[],"source_harness":"codex","source_row_index":380,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":64.73,"totalDurationS":191221.218,"totalInputTokens":125054108,"totalOutputTokens":1031431,"total_cost_usd":64.73,"total_input_tokens":125054108,"total_output_tokens":1031431,"total_runtime_s":191221.218,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.380952380952383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-153464e4dea13b8ebbcb","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-83471875135c750f4c3c00d1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3208955223880597,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":417,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[208];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":417,"metadata":{"avgScore":0.6068743663111191,"avg_score":0.6068743663111191,"avg_score_percent":60.687436631111915,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":33.74,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":62201618.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":448088.5,"passRate":0.3208955223880597,"pass_rate":0.3208955223880597,"pass_rate_percent":32.08955223880597,"passes":41,"runs":131,"runtime":116658,"source_flags":[],"source_harness":"codex","source_row_index":208,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":33.74,"totalDurationS":116658,"totalInputTokens":62201618.5,"totalOutputTokens":448088.5,"total_cost_usd":33.74,"total_input_tokens":62201618.5,"total_output_tokens":448088.5,"total_runtime_s":116658,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.08955223880597,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c649429a85762b4f312","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-809efc2ec8f8db45fa6647de","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":397,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[198];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":397,"metadata":{"avgScore":0.6595062733569761,"avg_score":0.6595062733569761,"avg_score_percent":65.9506273356976,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":47.32,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":84622600.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":590880,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":52,"runs":131,"runtime":130482.5,"source_flags":[],"source_harness":"codex","source_row_index":198,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":47.32,"totalDurationS":130482.5,"totalInputTokens":84622600.5,"totalOutputTokens":590880,"total_cost_usd":47.32,"total_input_tokens":84622600.5,"total_output_tokens":590880,"total_runtime_s":130482.5,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7770597d25acd08e13c7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7f02b08a4a256d380c9f103d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":291,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[145];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":291,"metadata":{"avgScore":0.1092923011990062,"avg_score":0.1092923011990062,"avg_score_percent":10.92923011990062,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":185.5,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":449191663,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1428521,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":73,"runtime":120135.5,"source_flags":[],"source_harness":"codex","source_row_index":145,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":185.5,"totalDurationS":120135.5,"totalInputTokens":449191663,"totalOutputTokens":1428521,"total_cost_usd":185.5,"total_input_tokens":449191663,"total_output_tokens":1428521,"total_runtime_s":120135.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bce7d5454e93a6777320","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7e33262d68f680ff230c2e4b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.43283582089552236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1217,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[608];split=unlicensed/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1217,"metadata":{"avgScore":0.7084704112498615,"avg_score":0.7084704112498615,"avg_score_percent":70.84704112498615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":83.8,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155958887.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1013526,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":57,"runs":133,"runtime":144963,"source_flags":[],"source_harness":"codex","source_row_index":608,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":83.8,"totalDurationS":144963,"totalInputTokens":155958887.5,"totalOutputTokens":1013526,"total_cost_usd":83.8,"total_input_tokens":155958887.5,"total_output_tokens":1013526,"total_runtime_s":144963,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.28358208955223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c3b43f8c3c35f0c3cff0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-73b046f1385667539683a0cf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1123,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[561];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1123,"metadata":{"avgScore":0.10630694231146019,"avg_score":0.10630694231146019,"avg_score_percent":10.63069423114602,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":87.9,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":204857419.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":845060.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":56,"runtime":75359.5,"source_flags":[],"source_harness":"codex","source_row_index":561,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":87.9,"totalDurationS":75359.5,"totalInputTokens":204857419.5,"totalOutputTokens":845060.5,"total_cost_usd":87.9,"total_input_tokens":204857419.5,"total_output_tokens":845060.5,"total_runtime_s":75359.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d4b7e8fa2b3b5803deca","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-72cee62b6d5aa82474310790","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29432624113475175,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":853,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[426];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":853,"metadata":{"avgScore":0.5252918479026842,"avg_score":0.5252918479026842,"avg_score_percent":52.529184790268424,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":499.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1147950590,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":5390942,"passRate":0.29432624113475175,"pass_rate":0.29432624113475175,"pass_rate_percent":29.432624113475175,"passes":81,"runs":277,"runtime":335753,"source_flags":[],"source_harness":"codex","source_row_index":426,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":499.67,"totalDurationS":335753,"totalInputTokens":1147950590,"totalOutputTokens":5390942,"total_cost_usd":499.67,"total_input_tokens":1147950590,"total_output_tokens":5390942,"total_runtime_s":335753,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.432624113475175,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-02b0b04700f461a955d8","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-704106bb8449d393213aefd8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":627,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[313];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":627,"metadata":{"avgScore":0.20454787385590675,"avg_score":0.20454787385590675,"avg_score_percent":20.454787385590674,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":28.04,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":53183956.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":200989.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":51418,"source_flags":[],"source_harness":"codex","source_row_index":313,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":28.04,"totalDurationS":51418,"totalInputTokens":53183956.5,"totalOutputTokens":200989.5,"total_cost_usd":28.04,"total_input_tokens":53183956.5,"total_output_tokens":200989.5,"total_runtime_s":51418,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f1e0d10950c473b9c3c3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6dedb701e19f518e788d04f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24468085106382978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1355,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[677];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1355,"metadata":{"avgScore":0.4438320139384285,"avg_score":0.4438320139384285,"avg_score_percent":44.38320139384285,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":111.31,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":213612416.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1450871,"passRate":0.24468085106382978,"pass_rate":0.24468085106382978,"pass_rate_percent":24.46808510638298,"passes":69,"runs":276,"runtime":284946.718,"source_flags":[],"source_harness":"codex","source_row_index":677,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":111.31,"totalDurationS":284946.718,"totalInputTokens":213612416.5,"totalOutputTokens":1450871,"total_cost_usd":111.31,"total_input_tokens":213612416.5,"total_output_tokens":1450871,"total_runtime_s":284946.718,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.46808510638298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f1f707f240c40e053812","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-60cad6ab6f6ff56335ab7717","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2198581560283688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":899,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[449];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":899,"metadata":{"avgScore":0.42499288919011063,"avg_score":0.42499288919011063,"avg_score_percent":42.499288919011065,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.42,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":153532999,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1050240.5,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":60,"runs":278,"runtime":250208.747,"source_flags":[],"source_harness":"codex","source_row_index":449,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":79.42,"totalDurationS":250208.747,"totalInputTokens":153532999,"totalOutputTokens":1050240.5,"total_cost_usd":79.42,"total_input_tokens":153532999,"total_output_tokens":1050240.5,"total_runtime_s":250208.747,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.98581560283688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f0fffbf3a56e351b4ab1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-557a5efa36c1be0ed7bd85fc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":961,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[480];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":961,"metadata":{"avgScore":0.4512691704888627,"avg_score":0.4512691704888627,"avg_score_percent":45.12691704888627,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":129.2,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":269660670.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1641035.5,"passRate":0.26,"pass_rate":0.26,"pass_rate_percent":26.0,"passes":26,"runs":99,"runtime":94791.5,"source_flags":[],"source_harness":"codex","source_row_index":480,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":129.2,"totalDurationS":94791.5,"totalInputTokens":269660670.5,"totalOutputTokens":1641035.5,"total_cost_usd":129.2,"total_input_tokens":269660670.5,"total_output_tokens":1641035.5,"total_runtime_s":94791.5,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e943dc24fbf63e87cb70","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5500fc5bccda45c5cde29e50","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1131,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[565];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1131,"metadata":{"avgScore":0.10286059924850519,"avg_score":0.10286059924850519,"avg_score_percent":10.286059924850518,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":32.32,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":71099685.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":415960,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":57,"runtime":74813.5,"source_flags":[],"source_harness":"codex","source_row_index":565,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":32.32,"totalDurationS":74813.5,"totalInputTokens":71099685.5,"totalOutputTokens":415960,"total_cost_usd":32.32,"total_input_tokens":71099685.5,"total_output_tokens":415960,"total_runtime_s":74813.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e88cb9085be13d7d9311","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-537c39e651f8b8fc33839420","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.20394736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":65,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[32];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":65,"metadata":{"avgScore":0.4024908249653849,"avg_score":0.4024908249653849,"avg_score_percent":40.24908249653849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.49,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":171679764,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1124088.5,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":60,"runs":300,"runtime":275061.747,"source_flags":[],"source_harness":"codex","source_row_index":32,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":89.49,"totalDurationS":275061.747,"totalInputTokens":171679764,"totalOutputTokens":1124088.5,"total_cost_usd":89.49,"total_input_tokens":171679764,"total_output_tokens":1124088.5,"total_runtime_s":275061.747,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.394736842105264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-50fa7541e1dafe7e59d7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-51a203c61bbe644cbd99d944","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.27631578947368424,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":499,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[249];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":499,"metadata":{"avgScore":0.484942264470657,"avg_score":0.484942264470657,"avg_score_percent":48.4942264470657,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":380.73,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":822499178,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":3817073,"passRate":0.27631578947368424,"pass_rate":0.27631578947368424,"pass_rate_percent":27.631578947368425,"passes":82,"runs":299,"runtime":314395.5,"source_flags":[],"source_harness":"codex","source_row_index":249,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":380.73,"totalDurationS":314395.5,"totalInputTokens":822499178,"totalOutputTokens":3817073,"total_cost_usd":380.73,"total_input_tokens":822499178,"total_output_tokens":3817073,"total_runtime_s":314395.5,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.631578947368425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-135edf7f7095b1bd562b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4dd3e406e619a27f438e97d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24285714285714285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":743,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[371];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":743,"metadata":{"avgScore":0.4688643981010808,"avg_score":0.4688643981010808,"avg_score_percent":46.88643981010808,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":144.1,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":297435267,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1832875,"passRate":0.24285714285714285,"pass_rate":0.24285714285714285,"pass_rate_percent":24.285714285714285,"passes":51,"runs":207,"runtime":167006.6195,"source_flags":[],"source_harness":"codex","source_row_index":371,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":144.1,"totalDurationS":167006.6195,"totalInputTokens":297435267,"totalOutputTokens":1832875,"total_cost_usd":144.1,"total_input_tokens":297435267,"total_output_tokens":1832875,"total_runtime_s":167006.6195,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.285714285714285,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-313bf882f4b7e5ba82ec","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4c84d239ba739b856023aeca","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":633,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[316];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":633,"metadata":{"avgScore":0.19554013940986384,"avg_score":0.19554013940986384,"avg_score_percent":19.554013940986383,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":16.92,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":31029753,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":114260.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":21,"runtime":25065.5,"source_flags":[],"source_harness":"codex","source_row_index":316,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":16.92,"totalDurationS":25065.5,"totalInputTokens":31029753,"totalOutputTokens":114260.5,"total_cost_usd":16.92,"total_input_tokens":31029753,"total_output_tokens":114260.5,"total_runtime_s":25065.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dde8a1b764acd061099f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4b02c4564e158fe113714e11","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.25,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":983,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[491];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":983,"metadata":{"avgScore":0.4148733134041199,"avg_score":0.4148733134041199,"avg_score_percent":41.48733134041199,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":91.53,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":177458632.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1109713,"passRate":0.25,"pass_rate":0.25,"pass_rate_percent":25.0,"passes":25,"runs":100,"runtime":73476,"source_flags":[],"source_harness":"codex","source_row_index":491,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":91.53,"totalDurationS":73476,"totalInputTokens":177458632.5,"totalOutputTokens":1109713,"total_cost_usd":91.53,"total_input_tokens":177458632.5,"total_output_tokens":1109713,"total_runtime_s":73476,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e0db23eacae8a33b34ff","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-498aa5a5cc8b56f3a1c9521e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1125,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[562];split=unlicensed/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1125,"metadata":{"avgScore":0.10511247258063579,"avg_score":0.10511247258063579,"avg_score_percent":10.511247258063579,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":23.24,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":51763892.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":279262.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":59,"runtime":73434,"source_flags":[],"source_harness":"codex","source_row_index":562,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":23.24,"totalDurationS":73434,"totalInputTokens":51763892.5,"totalOutputTokens":279262.5,"total_cost_usd":23.24,"total_input_tokens":51763892.5,"total_output_tokens":279262.5,"total_runtime_s":73434,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3a19826cc637c0906b0c","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-443f6ad4d2c2d062bef89a70","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":261,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[130];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":261,"metadata":{"avgScore":0.14914850377407576,"avg_score":0.14914850377407576,"avg_score_percent":14.914850377407577,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":264.49,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":678056488.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2240925.3333333335,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":2,"runs":74,"runtime":183287.5,"source_flags":[],"source_harness":"codex","source_row_index":130,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":264.49,"totalDurationS":183287.5,"totalInputTokens":678056488.1666666,"totalOutputTokens":2240925.3333333335,"total_cost_usd":264.49,"total_input_tokens":678056488.1666666,"total_output_tokens":2240925.3333333335,"total_runtime_s":183287.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4ceebcfeaa1516b9fbe7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3d647fad3b4fada0b0bdc56a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.43283582089552236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":381,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[190];split=full/near-term;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":381,"metadata":{"avgScore":0.7084704112498615,"avg_score":0.7084704112498615,"avg_score_percent":70.84704112498615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":83.8,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":155958887.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1013526,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":57,"runs":133,"runtime":144963,"source_flags":[],"source_harness":"codex","source_row_index":190,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":83.8,"totalDurationS":144963,"totalInputTokens":155958887.5,"totalOutputTokens":1013526,"total_cost_usd":83.8,"total_input_tokens":155958887.5,"total_output_tokens":1013526,"total_runtime_s":144963,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.28358208955223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ac9eb92f4b391de0d2ac","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3c12ee6898e35883af449837","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24468085106382978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":883,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[441];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":883,"metadata":{"avgScore":0.4438320139384285,"avg_score":0.4438320139384285,"avg_score_percent":44.38320139384285,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":111.31,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":213612416.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1450871,"passRate":0.24468085106382978,"pass_rate":0.24468085106382978,"pass_rate_percent":24.46808510638298,"passes":69,"runs":276,"runtime":284946.718,"source_flags":[],"source_harness":"codex","source_row_index":441,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":111.31,"totalDurationS":284946.718,"totalInputTokens":213612416.5,"totalOutputTokens":1450871,"total_cost_usd":111.31,"total_input_tokens":213612416.5,"total_output_tokens":1450871,"total_runtime_s":284946.718,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.46808510638298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ec13b7c7ae699ab83f0b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-33f3fa961cc6cc86114d6711","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2198581560283688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1371,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[685];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1371,"metadata":{"avgScore":0.42499288919011063,"avg_score":0.42499288919011063,"avg_score_percent":42.499288919011065,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.42,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":153532999,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1050240.5,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":60,"runs":278,"runtime":250208.747,"source_flags":[],"source_harness":"codex","source_row_index":685,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":79.42,"totalDurationS":250208.747,"totalInputTokens":153532999,"totalOutputTokens":1050240.5,"total_cost_usd":79.42,"total_input_tokens":153532999,"total_output_tokens":1050240.5,"total_runtime_s":250208.747,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.98581560283688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6882195e7a9d0dca72a5","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2c0936b0c3e0660b351e58ae","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2598684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":503,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[251];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":503,"metadata":{"avgScore":0.465716312064728,"avg_score":0.465716312064728,"avg_score_percent":46.5716312064728,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":261.53,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":533875323.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2793979.5,"passRate":0.2598684210526316,"pass_rate":0.2598684210526316,"pass_rate_percent":25.986842105263158,"passes":77,"runs":298,"runtime":354908.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":251,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":261.53,"totalDurationS":354908.11950000003,"totalInputTokens":533875323.5,"totalOutputTokens":2793979.5,"total_cost_usd":261.53,"total_input_tokens":533875323.5,"total_output_tokens":2793979.5,"total_runtime_s":354908.11950000003,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.986842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-846899d43b33cd6d3c59","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-220f5993422521d17d57b9d6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":635,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[317];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":635,"metadata":{"avgScore":0.19280668783242483,"avg_score":0.19280668783242483,"avg_score_percent":19.280668783242483,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.49,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":101601230,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":290736,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":53991.5,"source_flags":[],"source_harness":"codex","source_row_index":317,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":49.49,"totalDurationS":53991.5,"totalInputTokens":101601230,"totalOutputTokens":290736,"total_cost_usd":49.49,"total_input_tokens":101601230,"total_output_tokens":290736,"total_runtime_s":53991.5,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-081d47520eb9f3c0bb79","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-213a96723659590e92633d30","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":987,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[493];split=unlicensed/full-spectrum;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":987,"metadata":{"avgScore":0.39558362254691226,"avg_score":0.39558362254691226,"avg_score_percent":39.558362254691225,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":67.99,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":131879860,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":836993,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":21,"runs":99,"runtime":93679.6195,"source_flags":[],"source_harness":"codex","source_row_index":493,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":67.99,"totalDurationS":93679.6195,"totalInputTokens":131879860,"totalOutputTokens":836993,"total_cost_usd":67.99,"total_input_tokens":131879860,"total_output_tokens":836993,"total_runtime_s":93679.6195,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9bbb7c58f970c2ffb690","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1a36d2853d62528c19c97050","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22697368421052633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":517,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[258];split=full/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":517,"metadata":{"avgScore":0.4258635230185982,"avg_score":0.4258635230185982,"avg_score_percent":42.58635230185982,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":128.23,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":244642169.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":1565131.5,"passRate":0.22697368421052633,"pass_rate":0.22697368421052633,"pass_rate_percent":22.697368421052634,"passes":69,"runs":297,"runtime":310012.218,"source_flags":[],"source_harness":"codex","source_row_index":258,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":128.23,"totalDurationS":310012.218,"totalInputTokens":244642169.5,"totalOutputTokens":1565131.5,"total_cost_usd":128.23,"total_input_tokens":244642169.5,"total_output_tokens":1565131.5,"total_runtime_s":310012.218,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.697368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c2898a7b9bdccaad411d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-139e75eee6108a012ca3bd7d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":285,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[142];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":285,"metadata":{"avgScore":0.1122775555549231,"avg_score":0.1122775555549231,"avg_score_percent":11.22775555549231,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.18,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":65899280,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":335902,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":75,"runtime":86350,"source_flags":[],"source_harness":"codex","source_row_index":142,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":31.18,"totalDurationS":86350,"totalInputTokens":65899280,"totalOutputTokens":335902,"total_cost_usd":31.18,"total_input_tokens":65899280,"total_output_tokens":335902,"total_runtime_s":86350,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6d08c80309d9340b9559","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-10ab209ba267daf06730387c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2801418439716312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":867,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[433];split=unlicensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":867,"metadata":{"avgScore":0.48609115476187026,"avg_score":0.48609115476187026,"avg_score_percent":48.60911547618703,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":233.48,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":480691367,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2592990,"passRate":0.2801418439716312,"pass_rate":0.2801418439716312,"pass_rate_percent":28.01418439716312,"passes":77,"runs":276,"runtime":303490.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":433,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":233.48,"totalDurationS":303490.11950000003,"totalInputTokens":480691367,"totalOutputTokens":2592990,"total_cost_usd":233.48,"total_input_tokens":480691367,"total_output_tokens":2592990,"total_runtime_s":303490.11950000003,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.01418439716312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f69eb12e25ef35190cb4","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0e6d9429769fe5caad2204e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":279,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[139];split=full/last-exam;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":279,"metadata":{"avgScore":0.11983054733814444,"avg_score":0.11983054733814444,"avg_score_percent":11.983054733814443,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":105.65,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":236864478.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":967459,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":72,"runtime":102194.5,"source_flags":[],"source_harness":"codex","source_row_index":139,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":105.65,"totalDurationS":102194.5,"totalInputTokens":236864478.5,"totalOutputTokens":967459,"total_cost_usd":105.65,"total_input_tokens":236864478.5,"total_output_tokens":967459,"total_runtime_s":102194.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b856b7f1c166131efe68","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0e0c84c5fcd251dc04c9d508","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2619047619047619,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":731,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[365];split=linux_only;harness=codex;model=GPT-5.6-Terra;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":731,"metadata":{"avgScore":0.4936421663563447,"avg_score":0.4936421663563447,"avg_score_percent":49.36421663563447,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":234.69,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":528038236.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2578406,"passRate":0.2619047619047619,"pass_rate":0.2619047619047619,"pass_rate_percent":26.190476190476193,"passes":54,"runs":206,"runtime":176760,"source_flags":[],"source_harness":"codex","source_row_index":365,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":234.69,"totalDurationS":176760,"totalInputTokens":528038236.5,"totalOutputTokens":2578406,"total_cost_usd":234.69,"total_input_tokens":528038236.5,"total_output_tokens":2578406,"total_runtime_s":176760,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.190476190476193,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d6fd59dbd7ef54117677","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-04acad77d0f42e20c8a82999","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-max","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.29432624113475175,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1325,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[662];split=unlicensed/overall;harness=codex;model=GPT-5.6-Terra;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1325,"metadata":{"avgScore":0.5252918479026842,"avg_score":0.5252918479026842,"avg_score_percent":52.529184790268424,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":499.67,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-max","input_tokens":1147950590,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":5390942,"passRate":0.29432624113475175,"pass_rate":0.29432624113475175,"pass_rate_percent":29.432624113475175,"passes":81,"runs":277,"runtime":335753,"source_flags":[],"source_harness":"codex","source_row_index":662,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":499.67,"totalDurationS":335753,"totalInputTokens":1147950590,"totalOutputTokens":5390942,"total_cost_usd":499.67,"total_input_tokens":1147950590,"total_output_tokens":5390942,"total_runtime_s":335753,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.432624113475175,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-db4a18dd8e6e37a9b0df","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-03eb51219eb529a7699701d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":661,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[330];split=licensed;harness=codex;model=GPT-5.6-Terra;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":661,"metadata":{"avgScore":0.11405527444844456,"avg_score":0.11405527444844456,"avg_score_percent":11.405527444844456,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":10.07,"costSource":"list","cost_source":"list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":18146765,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":73848,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":22,"runtime":24853,"source_flags":[],"source_harness":"codex","source_row_index":330,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":10.07,"totalDurationS":24853,"totalInputTokens":18146765,"totalOutputTokens":73848,"total_cost_usd":10.07,"total_input_tokens":18146765,"total_output_tokens":73848,"total_runtime_s":24853,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-458496594621bde831ba","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0211dc49bc180c6210b9db4d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.6 Terra","modelRef":"GPT-5.6-Terra","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-high","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2598684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7688838b7684b3ae4b63f68c93ca4a31238131d2","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":31,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[15];split=full;harness=codex;model=GPT-5.6-Terra;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":31,"metadata":{"avgScore":0.465716312064728,"avg_score":0.465716312064728,"avg_score_percent":46.5716312064728,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":261.53,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":"reasoning-high","input_tokens":533875323.5,"metric_id":"pass_rate","metric_role":"primary","model":"GPT-5.6-Terra","output_tokens":2793979.5,"passRate":0.2598684210526316,"pass_rate":0.2598684210526316,"pass_rate_percent":25.986842105263158,"passes":77,"runs":298,"runtime":354908.11950000003,"source_flags":[],"source_harness":"codex","source_row_index":15,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":261.53,"totalDurationS":354908.11950000003,"totalInputTokens":533875323.5,"totalOutputTokens":2793979.5,"total_cost_usd":261.53,"total_input_tokens":533875323.5,"total_output_tokens":2793979.5,"total_runtime_s":354908.11950000003,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.986842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-08564b55eee6fee4dba7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-fc5db40705b62cb133ef480d","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4870141224330407,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":26,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[12];split=full;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":26,"metadata":{"avgScore":0.4870141224330407,"avg_score":0.4870141224330407,"avg_score_percent":48.70141224330407,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4340.24,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":2740460142,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":10436301,"passRate":0.2565789473684211,"pass_rate":0.2565789473684211,"pass_rate_percent":25.657894736842106,"passes":39,"runs":150,"runtime":251985,"source_flags":[],"source_harness":"claude_code","source_row_index":12,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":4340.24,"totalDurationS":251985,"totalInputTokens":2740460142,"totalOutputTokens":10436301,"total_cost_usd":4340.24,"total_input_tokens":2740460142,"total_output_tokens":10436301,"total_runtime_s":251985,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.70141224330407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-985f4afaf3007121e343","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e62cb31194b1292839b1a6ef","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.7112863360362098,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1216,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[607];split=unlicensed/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1216,"metadata":{"avgScore":0.7112863360362098,"avg_score":0.7112863360362098,"avg_score_percent":71.12863360362098,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1018.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":504938120,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3626698,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":67,"runtime":76455,"source_flags":[],"source_harness":"claude_code","source_row_index":607,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1018.09,"totalDurationS":76455,"totalInputTokens":504938120,"totalOutputTokens":3626698,"total_cost_usd":1018.09,"total_input_tokens":504938120,"total_output_tokens":3626698,"total_runtime_s":76455,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":71.12863360362098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5e0ca48965068ec74666","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-dc584149dff40c7f2e855725","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.43941419463910386,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1362,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[680];split=unlicensed/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1362,"metadata":{"avgScore":0.43941419463910386,"avg_score":0.43941419463910386,"avg_score_percent":43.941419463910385,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1828.62,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":745520634.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":8906113.5,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":35,"runs":146,"runtime":323705.5,"source_flags":[],"source_harness":"claude_code","source_row_index":680,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":1828.62,"totalDurationS":323705.5,"totalInputTokens":745520634.5,"totalOutputTokens":8906113.5,"total_cost_usd":1828.62,"total_input_tokens":745520634.5,"total_output_tokens":8906113.5,"total_runtime_s":323705.5,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.941419463910385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fc9bb5b49e358716119c","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-db63d2d33f3dbdbe9ec1daee","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07272727272727274,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":676,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[337];split=licensed;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":676,"metadata":{"avgScore":0.07272727272727274,"avg_score":0.07272727272727274,"avg_score_percent":7.272727272727273,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":486.11,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":134516414,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":529925,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":6,"runtime":58717,"source_flags":[],"source_harness":"claude_code","source_row_index":337,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":486.11,"totalDurationS":58717,"totalInputTokens":134516414,"totalOutputTokens":529925,"total_cost_usd":486.11,"total_input_tokens":134516414,"total_output_tokens":529925,"total_runtime_s":58717,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.272727272727273,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4b28278be71578589b40","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-dacb308f2f4610e3135820ae","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4200321393939394,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":146,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[72];split=full/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":146,"metadata":{"avgScore":0.4200321393939394,"avg_score":0.4200321393939394,"avg_score_percent":42.00321393939394,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1826.4,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1120840818,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3829080,"passRate":0.2545454545454545,"pass_rate":0.2545454545454545,"pass_rate_percent":25.454545454545453,"passes":14,"runs":55,"runtime":92024,"source_flags":[],"source_harness":"claude_code","source_row_index":72,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":1826.4,"totalDurationS":92024,"totalInputTokens":1120840818,"totalOutputTokens":3829080,"total_cost_usd":1826.4,"total_input_tokens":1120840818,"total_output_tokens":3829080,"total_runtime_s":92024,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.00321393939394,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0839e17ad309d14e1e38","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c209825b34c489bbc1b42ce9","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4044353533333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":986,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[492];split=unlicensed/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":986,"metadata":{"avgScore":0.4044353533333333,"avg_score":0.4044353533333333,"avg_score_percent":40.44353533333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1208.7,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":619603149,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3129503,"passRate":0.24,"pass_rate":0.24,"pass_rate_percent":24.0,"passes":12,"runs":50,"runtime":66288,"source_flags":[],"source_harness":"claude_code","source_row_index":492,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":1208.7,"totalDurationS":66288,"totalInputTokens":619603149,"totalOutputTokens":3129503,"total_cost_usd":1208.7,"total_input_tokens":619603149,"total_output_tokens":3129503,"total_runtime_s":66288,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.44353533333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1b7727ee57f0692b985","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b5cfddbd4ec1a08f48b60e58","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.412877641079695,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":54,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[26];split=full;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":54,"metadata":{"avgScore":0.412877641079695,"avg_score":0.412877641079695,"avg_score_percent":41.2877641079695,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2314.73,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":880037048.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":9436038.5,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":35,"runs":152,"runtime":382422.5,"source_flags":[],"source_harness":"claude_code","source_row_index":26,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":2314.73,"totalDurationS":382422.5,"totalInputTokens":880037048.5,"totalOutputTokens":9436038.5,"total_cost_usd":2314.73,"total_input_tokens":880037048.5,"total_output_tokens":9436038.5,"total_runtime_s":382422.5,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.2877641079695,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f1cbaf87d582e42c4f97","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-adf81a4204b484ab0d9514f6","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.16495248496656564,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":254,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[126];split=full/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":254,"metadata":{"avgScore":0.16495248496656564,"avg_score":0.16495248496656564,"avg_score_percent":16.495248496656565,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2091.31,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1585478623,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3847061,"passRate":0.07894736842105263,"pass_rate":0.07894736842105263,"pass_rate_percent":7.894736842105263,"passes":3,"runs":36,"runtime":108947,"source_flags":[],"source_harness":"claude_code","source_row_index":126,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":2091.31,"totalDurationS":108947,"totalInputTokens":1585478623,"totalOutputTokens":3847061,"total_cost_usd":2091.31,"total_input_tokens":1585478623,"total_output_tokens":3847061,"total_runtime_s":108947,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.495248496656565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-340771114f1b9347e7fe","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a8b9256d4ec8f5c57e18725d","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.06647333333333334,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1162,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[580];split=unlicensed/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1162,"metadata":{"avgScore":0.06647333333333334,"avg_score":0.06647333333333334,"avg_score_percent":6.647333333333334,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":402.44,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":121724035.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3125973.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":28,"runtime":140009.5,"source_flags":[],"source_harness":"claude_code","source_row_index":580,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":27,"tokens":null,"totalCostUsd":402.44,"totalDurationS":140009.5,"totalInputTokens":121724035.5,"totalOutputTokens":3125973.5,"total_cost_usd":402.44,"total_input_tokens":121724035.5,"total_output_tokens":3125973.5,"total_runtime_s":140009.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.647333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c27f1bb9c6ff91b55038","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8cf4e9b5ed70881506f4c519","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.4575886915040092,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":750,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[374];split=linux_only;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":750,"metadata":{"avgScore":0.4575886915040092,"avg_score":0.4575886915040092,"avg_score_percent":45.75886915040092,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":880.25,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":301356111.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":6923181.5,"passRate":0.22380952380952382,"pass_rate":0.22380952380952382,"pass_rate_percent":22.380952380952383,"passes":24,"runs":109,"runtime":250989,"source_flags":[],"source_harness":"claude_code","source_row_index":374,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":880.25,"totalDurationS":250989,"totalInputTokens":301356111.5,"totalOutputTokens":6923181.5,"total_cost_usd":880.25,"total_input_tokens":301356111.5,"total_output_tokens":6923181.5,"total_runtime_s":250989,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.75886915040092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0af33e56f59412947a65","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7cd2ced0be2c6c41c9da4aa0","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.43941419463910386,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":890,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[444];split=unlicensed;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":890,"metadata":{"avgScore":0.43941419463910386,"avg_score":0.43941419463910386,"avg_score_percent":43.941419463910385,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1828.62,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":745520634.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":8906113.5,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":35,"runs":146,"runtime":323705.5,"source_flags":[],"source_harness":"claude_code","source_row_index":444,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":1828.62,"totalDurationS":323705.5,"totalInputTokens":745520634.5,"totalOutputTokens":8906113.5,"total_cost_usd":1828.62,"total_input_tokens":745520634.5,"total_output_tokens":8906113.5,"total_runtime_s":323705.5,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.941419463910385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-921d5f9c68c4045c5380","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7951d220da5f9a2b3a9c1a47","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4845098377095127,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":738,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[368];split=linux_only;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":738,"metadata":{"avgScore":0.4845098377095127,"avg_score":0.4845098377095127,"avg_score_percent":48.450983770951275,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1662.85,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":900239978,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":6378495,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":25,"runs":105,"runtime":124476,"source_flags":[],"source_harness":"claude_code","source_row_index":368,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":1662.85,"totalDurationS":124476,"totalInputTokens":900239978,"totalOutputTokens":6378495,"total_cost_usd":1662.85,"total_input_tokens":900239978,"total_output_tokens":6378495,"total_runtime_s":124476,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.450983770951275,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aedd581b714ede4035b6","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7881459fc99581d460983a7a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.16393594595273656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1092,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[545];split=unlicensed/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1092,"metadata":{"avgScore":0.16393594595273656,"avg_score":0.16393594595273656,"avg_score_percent":16.393594595273655,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1087.64,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":734827171,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":2889009,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":29,"runtime":63512,"source_flags":[],"source_harness":"claude_code","source_row_index":545,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":1087.64,"totalDurationS":63512,"totalInputTokens":734827171,"totalOutputTokens":2889009,"total_cost_usd":1087.64,"total_input_tokens":734827171,"total_output_tokens":2889009,"total_runtime_s":63512,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.393594595273655,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3d365d51d1faeea21ce0","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-648c8534aaf07f05a2c409c0","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.412877641079695,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":526,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[262];split=full/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":526,"metadata":{"avgScore":0.412877641079695,"avg_score":0.412877641079695,"avg_score_percent":41.2877641079695,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2314.73,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":880037048.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":9436038.5,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":35,"runs":152,"runtime":382422.5,"source_flags":[],"source_harness":"claude_code","source_row_index":262,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":2314.73,"totalDurationS":382422.5,"totalInputTokens":880037048.5,"totalOutputTokens":9436038.5,"total_cost_usd":2314.73,"total_input_tokens":880037048.5,"total_output_tokens":9436038.5,"total_runtime_s":382422.5,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.2877641079695,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-530023b34acfef3e0851","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-6311802c2a90c3322c991aee","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.6414472929875101,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1236,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[617];split=unlicensed/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1236,"metadata":{"avgScore":0.6414472929875101,"avg_score":0.6414472929875101,"avg_score_percent":64.14472929875102,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":947.17,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":341718065,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3125447.5,"passRate":0.35074626865671643,"pass_rate":0.35074626865671643,"pass_rate_percent":35.07462686567165,"passes":24,"runs":72,"runtime":114531,"source_flags":[],"source_harness":"claude_code","source_row_index":617,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":947.17,"totalDurationS":114531,"totalInputTokens":341718065,"totalOutputTokens":3125447.5,"total_cost_usd":947.17,"total_input_tokens":341718065,"total_output_tokens":3125447.5,"total_runtime_s":114531,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":64.14472929875102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-049611f59448512273cd","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-507b3168305944757e0e8714","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.052478947368421056,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":328,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[163];split=full/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":328,"metadata":{"avgScore":0.052478947368421056,"avg_score":0.052478947368421056,"avg_score_percent":5.247894736842105,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":838.46,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":231247542.5,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3319023.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":188112.5,"source_flags":[],"source_harness":"claude_code","source_row_index":163,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":30,"tokens":null,"totalCostUsd":838.46,"totalDurationS":188112.5,"totalInputTokens":231247542.5,"totalOutputTokens":3319023.5,"total_cost_usd":838.46,"total_input_tokens":231247542.5,"total_output_tokens":3319023.5,"total_runtime_s":188112.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":5.247894736842105,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-920856601ed297814654","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-493a0b4b38d0d0f76fe6eb71","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5020995075154241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":864,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[431];split=unlicensed;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":864,"metadata":{"avgScore":0.5020995075154241,"avg_score":0.5020995075154241,"avg_score_percent":50.20995075154241,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3187.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1779144434,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":9163045,"passRate":0.2695035460992908,"pass_rate":0.2695035460992908,"pass_rate_percent":26.95035460992908,"passes":38,"runs":140,"runtime":196439,"source_flags":[],"source_harness":"claude_code","source_row_index":431,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":3187.09,"totalDurationS":196439,"totalInputTokens":1779144434,"totalOutputTokens":9163045,"total_cost_usd":3187.09,"total_input_tokens":1779144434,"total_output_tokens":9163045,"total_runtime_s":196439,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.20995075154241,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1c2e1b8ecf643b4ddd5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2e99445131cd50bb13ee40e6","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5020995075154241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1336,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[667];split=unlicensed/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1336,"metadata":{"avgScore":0.5020995075154241,"avg_score":0.5020995075154241,"avg_score_percent":50.20995075154241,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3187.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1779144434,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":9163045,"passRate":0.2695035460992908,"pass_rate":0.2695035460992908,"pass_rate_percent":26.95035460992908,"passes":38,"runs":140,"runtime":196439,"source_flags":[],"source_harness":"claude_code","source_row_index":667,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":3187.09,"totalDurationS":196439,"totalInputTokens":1779144434,"totalOutputTokens":9163045,"total_cost_usd":3187.09,"total_input_tokens":1779144434,"total_output_tokens":9163045,"total_runtime_s":196439,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.20995075154241,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1d62801c6d6ee9c251e","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-28d7c8a0e773b16159afb1ff","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.7112863360362098,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":380,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[189];split=full/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":380,"metadata":{"avgScore":0.7112863360362098,"avg_score":0.7112863360362098,"avg_score_percent":71.12863360362098,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1018.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":504938120,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3626698,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":67,"runtime":76455,"source_flags":[],"source_harness":"claude_code","source_row_index":189,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1018.09,"totalDurationS":76455,"totalInputTokens":504938120,"totalOutputTokens":3626698,"total_cost_usd":1018.09,"total_input_tokens":504938120,"total_output_tokens":3626698,"total_runtime_s":76455,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":71.12863360362098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-35b1a95b6b095210c54f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-264ca74cbc9a8e04933ff307","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.35374968752637187,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":176,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[87];split=full/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":176,"metadata":{"avgScore":0.35374968752637187,"avg_score":0.35374968752637187,"avg_score_percent":35.37496875263719,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":608.41,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":327686352,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3672954.5,"passRate":0.20909090909090908,"pass_rate":0.20909090909090908,"pass_rate_percent":20.909090909090907,"passes":12,"runs":57,"runtime":115842,"source_flags":[],"source_harness":"claude_code","source_row_index":87,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":608.41,"totalDurationS":115842,"totalInputTokens":327686352,"totalOutputTokens":3672954.5,"total_cost_usd":608.41,"total_input_tokens":327686352,"total_output_tokens":3672954.5,"total_runtime_s":115842,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.37496875263719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9754ef2c025b3869a9d7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-23979b88953963b0f01f7df4","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.37312465627900904,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1006,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[502];split=unlicensed/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1006,"metadata":{"avgScore":0.37312465627900904,"avg_score":0.37312465627900904,"avg_score_percent":37.312465627900906,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":555.24,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":301449420,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3311822.5,"passRate":0.23,"pass_rate":0.23,"pass_rate_percent":23.0,"passes":12,"runs":52,"runtime":86760,"source_flags":[],"source_harness":"claude_code","source_row_index":502,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":555.24,"totalDurationS":86760,"totalInputTokens":301449420,"totalOutputTokens":3311822.5,"total_cost_usd":555.24,"total_input_tokens":301449420,"total_output_tokens":3311822.5,"total_runtime_s":86760,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.312465627900906,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-135906311067543096f7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2360d3dffb9156625513dff2","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.6414472929875101,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":400,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[199];split=full/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":400,"metadata":{"avgScore":0.6414472929875101,"avg_score":0.6414472929875101,"avg_score_percent":64.14472929875102,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":947.17,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":341718065,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":3125447.5,"passRate":0.35074626865671643,"pass_rate":0.35074626865671643,"pass_rate_percent":35.07462686567165,"passes":24,"runs":72,"runtime":114531,"source_flags":[],"source_harness":"claude_code","source_row_index":199,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":947.17,"totalDurationS":114531,"totalInputTokens":341718065,"totalOutputTokens":3125447.5,"total_cost_usd":947.17,"total_input_tokens":341718065,"total_output_tokens":3125447.5,"total_runtime_s":114531,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":64.14472929875102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-52d9f37a04c82033e4ff","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-218e4d132902de1050d7c9a1","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2936469136497635,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":598,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[298];split=licensed;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":598,"metadata":{"avgScore":0.2936469136497635,"avg_score":0.2936469136497635,"avg_score_percent":29.36469136497635,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1153.16,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":961315708,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":1273256,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":10,"runtime":55546,"source_flags":[],"source_harness":"claude_code","source_row_index":298,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":1153.16,"totalDurationS":55546,"totalInputTokens":961315708,"totalOutputTokens":1273256,"total_cost_usd":1153.16,"total_input_tokens":961315708,"total_output_tokens":1273256,"total_runtime_s":55546,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.36469136497635,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8696c0f42e27e873a8b0","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-091a40ab5087035956ac297e","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4870141224330407,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bd7563a7bd198a593fcfff5360082a7b0b40189","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":498,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[248];split=full/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":498,"metadata":{"avgScore":0.4870141224330407,"avg_score":0.4870141224330407,"avg_score_percent":48.70141224330407,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4340.24,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":2740460142,"metric_id":"avg_score","metric_role":"secondary","model":"anthropic-claude-fable-5","output_tokens":10436301,"passRate":0.2565789473684211,"pass_rate":0.2565789473684211,"pass_rate_percent":25.657894736842106,"passes":39,"runs":150,"runtime":251985,"source_flags":[],"source_harness":"claude_code","source_row_index":248,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":4340.24,"totalDurationS":251985,"totalInputTokens":2740460142,"totalOutputTokens":10436301,"total_cost_usd":4340.24,"total_input_tokens":2740460142,"total_output_tokens":10436301,"total_runtime_s":251985,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.70141224330407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f23555b8796467b78014","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f22bc37454a1f03e253b069c","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":327,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[163];split=full/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":327,"metadata":{"avgScore":0.052478947368421056,"avg_score":0.052478947368421056,"avg_score_percent":5.247894736842105,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":838.46,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":231247542.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3319023.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":188112.5,"source_flags":[],"source_harness":"claude_code","source_row_index":163,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":30,"tokens":null,"totalCostUsd":838.46,"totalDurationS":188112.5,"totalInputTokens":231247542.5,"totalOutputTokens":3319023.5,"total_cost_usd":838.46,"total_input_tokens":231247542.5,"total_output_tokens":3319023.5,"total_runtime_s":188112.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-59d26884b508cae47e5d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ef9d6c744650adde794dcaa9","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.24,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":985,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[492];split=unlicensed/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":985,"metadata":{"avgScore":0.4044353533333333,"avg_score":0.4044353533333333,"avg_score_percent":40.44353533333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1208.7,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":619603149,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3129503,"passRate":0.24,"pass_rate":0.24,"pass_rate_percent":24.0,"passes":12,"runs":50,"runtime":66288,"source_flags":[],"source_harness":"claude_code","source_row_index":492,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":1208.7,"totalDurationS":66288,"totalInputTokens":619603149,"totalOutputTokens":3129503,"total_cost_usd":1208.7,"total_input_tokens":619603149,"total_output_tokens":3129503,"total_runtime_s":66288,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-af01a4f9fc83ad3fc941","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ea78c40e4dbff3a51bbca76e","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":675,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[337];split=licensed;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":675,"metadata":{"avgScore":0.07272727272727274,"avg_score":0.07272727272727274,"avg_score_percent":7.272727272727273,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":486.11,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":134516414,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":529925,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":6,"runtime":58717,"source_flags":[],"source_harness":"claude_code","source_row_index":337,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":486.11,"totalDurationS":58717,"totalInputTokens":134516414,"totalOutputTokens":529925,"total_cost_usd":486.11,"total_input_tokens":134516414,"total_output_tokens":529925,"total_runtime_s":58717,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4ae64b85b28a8e8f3077","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ea08c5397aa36f60cbf3e3e4","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2545454545454545,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":145,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[72];split=full/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":145,"metadata":{"avgScore":0.4200321393939394,"avg_score":0.4200321393939394,"avg_score_percent":42.00321393939394,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1826.4,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1120840818,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3829080,"passRate":0.2545454545454545,"pass_rate":0.2545454545454545,"pass_rate_percent":25.454545454545453,"passes":14,"runs":55,"runtime":92024,"source_flags":[],"source_harness":"claude_code","source_row_index":72,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":1826.4,"totalDurationS":92024,"totalInputTokens":1120840818,"totalOutputTokens":3829080,"total_cost_usd":1826.4,"total_input_tokens":1120840818,"total_output_tokens":3829080,"total_runtime_s":92024,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.454545454545453,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0601426179ac0ca44f73","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e75ffb0d5d696c9be27a2bbf","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1361,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[680];split=unlicensed/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1361,"metadata":{"avgScore":0.43941419463910386,"avg_score":0.43941419463910386,"avg_score_percent":43.941419463910385,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1828.62,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":745520634.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":8906113.5,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":35,"runs":146,"runtime":323705.5,"source_flags":[],"source_harness":"claude_code","source_row_index":680,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":1828.62,"totalDurationS":323705.5,"totalInputTokens":745520634.5,"totalOutputTokens":8906113.5,"total_cost_usd":1828.62,"total_input_tokens":745520634.5,"total_output_tokens":8906113.5,"total_runtime_s":323705.5,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b35698f8b17a8f8aaa66","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-daa59ad914f70d7f944df34a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":525,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[262];split=full/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":525,"metadata":{"avgScore":0.412877641079695,"avg_score":0.412877641079695,"avg_score_percent":41.2877641079695,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2314.73,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":880037048.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":9436038.5,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":35,"runs":152,"runtime":382422.5,"source_flags":[],"source_harness":"claude_code","source_row_index":262,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":2314.73,"totalDurationS":382422.5,"totalInputTokens":880037048.5,"totalOutputTokens":9436038.5,"total_cost_usd":2314.73,"total_input_tokens":880037048.5,"total_output_tokens":9436038.5,"total_runtime_s":382422.5,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.36842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7c1e73bf5c7f29b3ef19","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c3ef91575df831219e67b18e","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2565789473684211,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":497,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[248];split=full/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":497,"metadata":{"avgScore":0.4870141224330407,"avg_score":0.4870141224330407,"avg_score_percent":48.70141224330407,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4340.24,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":2740460142,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":10436301,"passRate":0.2565789473684211,"pass_rate":0.2565789473684211,"pass_rate_percent":25.657894736842106,"passes":39,"runs":150,"runtime":251985,"source_flags":[],"source_harness":"claude_code","source_row_index":248,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":4340.24,"totalDurationS":251985,"totalInputTokens":2740460142,"totalOutputTokens":10436301,"total_cost_usd":4340.24,"total_input_tokens":2740460142,"total_output_tokens":10436301,"total_runtime_s":251985,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.657894736842106,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fb8101d498be04e1bd63","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-bb88cb8af3e00615cc07311c","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.07894736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":253,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[126];split=full/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":253,"metadata":{"avgScore":0.16495248496656564,"avg_score":0.16495248496656564,"avg_score_percent":16.495248496656565,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2091.31,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1585478623,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3847061,"passRate":0.07894736842105263,"pass_rate":0.07894736842105263,"pass_rate_percent":7.894736842105263,"passes":3,"runs":36,"runtime":108947,"source_flags":[],"source_harness":"claude_code","source_row_index":126,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":2091.31,"totalDurationS":108947,"totalInputTokens":1585478623,"totalOutputTokens":3847061,"total_cost_usd":2091.31,"total_input_tokens":1585478623,"total_output_tokens":3847061,"total_runtime_s":108947,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.894736842105263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4d9b9b78521e0615e7f4","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-adc655fbfe8573bf1ba9e53f","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2695035460992908,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1335,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[667];split=unlicensed/overall;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1335,"metadata":{"avgScore":0.5020995075154241,"avg_score":0.5020995075154241,"avg_score_percent":50.20995075154241,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3187.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1779144434,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":9163045,"passRate":0.2695035460992908,"pass_rate":0.2695035460992908,"pass_rate_percent":26.95035460992908,"passes":38,"runs":140,"runtime":196439,"source_flags":[],"source_harness":"claude_code","source_row_index":667,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":3187.09,"totalDurationS":196439,"totalInputTokens":1779144434,"totalOutputTokens":9163045,"total_cost_usd":3187.09,"total_input_tokens":1779144434,"total_output_tokens":9163045,"total_runtime_s":196439,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.95035460992908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a5c7b4120941be3c8a6","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-913d2dbcf9c3bf2f26184cf1","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":53,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[26];split=full;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":53,"metadata":{"avgScore":0.412877641079695,"avg_score":0.412877641079695,"avg_score_percent":41.2877641079695,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2314.73,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":880037048.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":9436038.5,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":35,"runs":152,"runtime":382422.5,"source_flags":[],"source_harness":"claude_code","source_row_index":26,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":2314.73,"totalDurationS":382422.5,"totalInputTokens":880037048.5,"totalOutputTokens":9436038.5,"total_cost_usd":2314.73,"total_input_tokens":880037048.5,"total_output_tokens":9436038.5,"total_runtime_s":382422.5,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.36842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-23856d68b2625bcd3df8","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8f191c3dd1dd717d19299a33","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2565789473684211,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":25,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[12];split=full;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":25,"metadata":{"avgScore":0.4870141224330407,"avg_score":0.4870141224330407,"avg_score_percent":48.70141224330407,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4340.24,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":2740460142,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":10436301,"passRate":0.2565789473684211,"pass_rate":0.2565789473684211,"pass_rate_percent":25.657894736842106,"passes":39,"runs":150,"runtime":251985,"source_flags":[],"source_harness":"claude_code","source_row_index":12,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":4340.24,"totalDurationS":251985,"totalInputTokens":2740460142,"totalOutputTokens":10436301,"total_cost_usd":4340.24,"total_input_tokens":2740460142,"total_output_tokens":10436301,"total_runtime_s":251985,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.657894736842106,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3189d2f120ba645985e4","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-89ec55f67a3e6e7e3a66fdcb","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2695035460992908,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":863,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[431];split=unlicensed;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":863,"metadata":{"avgScore":0.5020995075154241,"avg_score":0.5020995075154241,"avg_score_percent":50.20995075154241,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3187.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":1779144434,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":9163045,"passRate":0.2695035460992908,"pass_rate":0.2695035460992908,"pass_rate_percent":26.95035460992908,"passes":38,"runs":140,"runtime":196439,"source_flags":[],"source_harness":"claude_code","source_row_index":431,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":3187.09,"totalDurationS":196439,"totalInputTokens":1779144434,"totalOutputTokens":9163045,"total_cost_usd":3187.09,"total_input_tokens":1779144434,"total_output_tokens":9163045,"total_runtime_s":196439,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.95035460992908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7ec5a17d39dbdcfb712f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8385d5f48ce1c6482c0ddc7f","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.20909090909090908,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":175,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[87];split=full/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":175,"metadata":{"avgScore":0.35374968752637187,"avg_score":0.35374968752637187,"avg_score_percent":35.37496875263719,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":608.41,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":327686352,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3672954.5,"passRate":0.20909090909090908,"pass_rate":0.20909090909090908,"pass_rate_percent":20.909090909090907,"passes":12,"runs":57,"runtime":115842,"source_flags":[],"source_harness":"claude_code","source_row_index":87,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":608.41,"totalDurationS":115842,"totalInputTokens":327686352,"totalOutputTokens":3672954.5,"total_cost_usd":608.41,"total_input_tokens":327686352,"total_output_tokens":3672954.5,"total_runtime_s":115842,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.909090909090907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-996c3c87fbc778d4720e","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-6ed23bc2eb2eb873a011da95","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":597,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[298];split=licensed;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":597,"metadata":{"avgScore":0.2936469136497635,"avg_score":0.2936469136497635,"avg_score_percent":29.36469136497635,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1153.16,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":961315708,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":1273256,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":10,"runtime":55546,"source_flags":[],"source_harness":"claude_code","source_row_index":298,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":1153.16,"totalDurationS":55546,"totalInputTokens":961315708,"totalOutputTokens":1273256,"total_cost_usd":1153.16,"total_input_tokens":961315708,"total_output_tokens":1273256,"total_runtime_s":55546,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9758fa79ae0b3daa8da8","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5db3ff1096881ef3e76cc8f8","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1161,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[580];split=unlicensed/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1161,"metadata":{"avgScore":0.06647333333333334,"avg_score":0.06647333333333334,"avg_score_percent":6.647333333333334,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":402.44,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":121724035.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3125973.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":28,"runtime":140009.5,"source_flags":[],"source_harness":"claude_code","source_row_index":580,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":27,"tokens":null,"totalCostUsd":402.44,"totalDurationS":140009.5,"totalInputTokens":121724035.5,"totalOutputTokens":3125973.5,"total_cost_usd":402.44,"total_input_tokens":121724035.5,"total_output_tokens":3125973.5,"total_runtime_s":140009.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-48ef0b9061edcd799599","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-55e3add333bbca37eec9494d","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1091,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[545];split=unlicensed/last-exam;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1091,"metadata":{"avgScore":0.16393594595273656,"avg_score":0.16393594595273656,"avg_score_percent":16.393594595273655,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1087.64,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":734827171,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":2889009,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":29,"runtime":63512,"source_flags":[],"source_harness":"claude_code","source_row_index":545,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":1087.64,"totalDurationS":63512,"totalInputTokens":734827171,"totalOutputTokens":2889009,"total_cost_usd":1087.64,"total_input_tokens":734827171,"total_output_tokens":2889009,"total_runtime_s":63512,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cca51e9cdbdd9a508b95","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-49e576eb38f9aa8bc33cbc9c","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.35074626865671643,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":399,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[199];split=full/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":399,"metadata":{"avgScore":0.6414472929875101,"avg_score":0.6414472929875101,"avg_score_percent":64.14472929875102,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":947.17,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":341718065,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3125447.5,"passRate":0.35074626865671643,"pass_rate":0.35074626865671643,"pass_rate_percent":35.07462686567165,"passes":24,"runs":72,"runtime":114531,"source_flags":[],"source_harness":"claude_code","source_row_index":199,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":947.17,"totalDurationS":114531,"totalInputTokens":341718065,"totalOutputTokens":3125447.5,"total_cost_usd":947.17,"total_input_tokens":341718065,"total_output_tokens":3125447.5,"total_runtime_s":114531,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.07462686567165,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1529ecf659228ed14fc","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4275edd6a1c6a00447702fac","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.22380952380952382,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":749,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[374];split=linux_only;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":749,"metadata":{"avgScore":0.4575886915040092,"avg_score":0.4575886915040092,"avg_score_percent":45.75886915040092,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":880.25,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":301356111.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":6923181.5,"passRate":0.22380952380952382,"pass_rate":0.22380952380952382,"pass_rate_percent":22.380952380952383,"passes":24,"runs":109,"runtime":250989,"source_flags":[],"source_harness":"claude_code","source_row_index":374,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":880.25,"totalDurationS":250989,"totalInputTokens":301356111.5,"totalOutputTokens":6923181.5,"total_cost_usd":880.25,"total_input_tokens":301356111.5,"total_output_tokens":6923181.5,"total_runtime_s":250989,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.380952380952383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-59775944d0feb95f47ff","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-393ecf8640686598f8510bb7","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.373134328358209,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":379,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[189];split=full/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":379,"metadata":{"avgScore":0.7112863360362098,"avg_score":0.7112863360362098,"avg_score_percent":71.12863360362098,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1018.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":504938120,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3626698,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":67,"runtime":76455,"source_flags":[],"source_harness":"claude_code","source_row_index":189,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1018.09,"totalDurationS":76455,"totalInputTokens":504938120,"totalOutputTokens":3626698,"total_cost_usd":1018.09,"total_input_tokens":504938120,"total_output_tokens":3626698,"total_runtime_s":76455,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.3134328358209,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d305faa89f99e99b9378","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1b5c05ee6c0bd5f7ae2b84c5","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":889,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[444];split=unlicensed;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":889,"metadata":{"avgScore":0.43941419463910386,"avg_score":0.43941419463910386,"avg_score_percent":43.941419463910385,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1828.62,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":745520634.5,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":8906113.5,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":35,"runs":146,"runtime":323705.5,"source_flags":[],"source_harness":"claude_code","source_row_index":444,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":1828.62,"totalDurationS":323705.5,"totalInputTokens":745520634.5,"totalOutputTokens":8906113.5,"total_cost_usd":1828.62,"total_input_tokens":745520634.5,"total_output_tokens":8906113.5,"total_runtime_s":323705.5,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7a35d9a04c2eb72d088d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-10e35d8300079110fe6082f8","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.23809523809523808,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":737,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[368];split=linux_only;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":737,"metadata":{"avgScore":0.4845098377095127,"avg_score":0.4845098377095127,"avg_score_percent":48.450983770951275,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1662.85,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":900239978,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":6378495,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":25,"runs":105,"runtime":124476,"source_flags":[],"source_harness":"claude_code","source_row_index":368,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":1662.85,"totalDurationS":124476,"totalInputTokens":900239978,"totalOutputTokens":6378495,"total_cost_usd":1662.85,"total_input_tokens":900239978,"total_output_tokens":6378495,"total_runtime_s":124476,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.809523809523807,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6c7677a786dcaa12e156","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-10689956bfacf31d2b02855e","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.23,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1005,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[502];split=unlicensed/full-spectrum;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1005,"metadata":{"avgScore":0.37312465627900904,"avg_score":0.37312465627900904,"avg_score_percent":37.312465627900906,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":555.24,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":301449420,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3311822.5,"passRate":0.23,"pass_rate":0.23,"pass_rate_percent":23.0,"passes":12,"runs":52,"runtime":86760,"source_flags":[],"source_harness":"claude_code","source_row_index":502,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":555.24,"totalDurationS":86760,"totalInputTokens":301449420,"totalOutputTokens":3311822.5,"total_cost_usd":555.24,"total_input_tokens":301449420,"total_output_tokens":3311822.5,"total_runtime_s":86760,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0ca3b469a83415ae14d5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0a7ef4ada9d06512216e5dfe","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.35074626865671643,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1235,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[617];split=unlicensed/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1235,"metadata":{"avgScore":0.6414472929875101,"avg_score":0.6414472929875101,"avg_score_percent":64.14472929875102,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":947.17,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":341718065,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3125447.5,"passRate":0.35074626865671643,"pass_rate":0.35074626865671643,"pass_rate_percent":35.07462686567165,"passes":24,"runs":72,"runtime":114531,"source_flags":[],"source_harness":"claude_code","source_row_index":617,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":947.17,"totalDurationS":114531,"totalInputTokens":341718065,"totalOutputTokens":3125447.5,"total_cost_usd":947.17,"total_input_tokens":341718065,"total_output_tokens":3125447.5,"total_runtime_s":114531,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.07462686567165,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a318857c95fa1d6ace3","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-09bbad213c828a339c6199e9","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Fable 5","modelRef":"anthropic-claude-fable-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-xhigh","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.373134328358209,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89a94910fbad2f09f3d3794332a63b055ce236bc","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1215,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[607];split=unlicensed/near-term;harness=claude_code;model=anthropic-claude-fable-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1215,"metadata":{"avgScore":0.7112863360362098,"avg_score":0.7112863360362098,"avg_score_percent":71.12863360362098,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1018.09,"costSource":"official","cost_source":"official","harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-xhigh","input_tokens":504938120,"metric_id":"pass_rate","metric_role":"primary","model":"anthropic-claude-fable-5","output_tokens":3626698,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":67,"runtime":76455,"source_flags":[],"source_harness":"claude_code","source_row_index":607,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1018.09,"totalDurationS":76455,"totalInputTokens":504938120,"totalOutputTokens":3626698,"total_cost_usd":1018.09,"total_input_tokens":504938120,"total_output_tokens":3626698,"total_runtime_s":76455,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.3134328358209,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-78e86970acef878dd4b7","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fd8bf989d55c7b7e19c72803","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6045069241062173,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1256,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[627];split=unlicensed/near-term;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1256,"metadata":{"avgScore":0.6045069241062173,"avg_score":0.6045069241062173,"avg_score_percent":60.450692410621734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":311.72,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":359152755,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1944556,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":74416.76165142255,"source_flags":[],"source_harness":"ale_claw","source_row_index":627,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":311.72,"totalDurationS":74416.76165142255,"totalInputTokens":359152755,"totalOutputTokens":1944556,"total_cost_usd":311.72,"total_input_tokens":359152755,"total_output_tokens":1944556,"total_runtime_s":74416.76165142255,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":60.450692410621734,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-167b3f44ab544a5b10b7","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fb6c9f79a9a54c39d21e4a45","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6171275522572685,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1250,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[624];split=unlicensed/near-term;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1250,"metadata":{"avgScore":0.6171275522572685,"avg_score":0.6171275522572685,"avg_score_percent":61.712755225726845,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":737.55,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":176244296.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1470123,"passRate":0.291044776119403,"pass_rate":0.291044776119403,"pass_rate_percent":29.1044776119403,"passes":21,"runs":69,"runtime":57944.460609696805,"source_flags":[],"source_harness":"droid","source_row_index":624,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":737.55,"totalDurationS":57944.460609696805,"totalInputTokens":176244296.5,"totalOutputTokens":1470123,"total_cost_usd":737.55,"total_input_tokens":176244296.5,"total_output_tokens":1470123,"total_runtime_s":57944.460609696805,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":61.712755225726845,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-56380f1a6390f96eb1ed","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f710de4f65d5f894e9f14a49","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10856285924276156,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":232,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[115];split=full/full-spectrum;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":232,"metadata":{"avgScore":0.10856285924276156,"avg_score":0.10856285924276156,"avg_score_percent":10.856285924276156,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":166.81,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":39918384,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":472069,"passRate":0.03636363636363636,"pass_rate":0.03636363636363636,"pass_rate_percent":3.6363636363636362,"passes":2,"runs":20,"runtime":11256.058063949225,"source_flags":[],"source_harness":"droid","source_row_index":115,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":20,"tokens":null,"totalCostUsd":166.81,"totalDurationS":11256.058063949225,"totalInputTokens":39918384,"totalOutputTokens":472069,"total_cost_usd":166.81,"total_input_tokens":39918384,"total_output_tokens":472069,"total_runtime_s":11256.058063949225,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.856285924276156,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-048217f21b35e9f68caf","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f1f4935fea7cfcdac6502e94","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.3890574125005675,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":790,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[394];split=linux_only;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":790,"metadata":{"avgScore":0.3890574125005675,"avg_score":0.3890574125005675,"avg_score_percent":38.90574125005675,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1164.8,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":280978346,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2723233,"passRate":0.1523809523809524,"pass_rate":0.1523809523809524,"pass_rate_percent":15.238095238095239,"passes":16,"runs":96,"runtime":92839.4148778622,"source_flags":[],"source_harness":"claude_code","source_row_index":394,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":96,"tokens":null,"totalCostUsd":1164.8,"totalDurationS":92839.4148778622,"totalInputTokens":280978346,"totalOutputTokens":2723233,"total_cost_usd":1164.8,"total_input_tokens":280978346,"total_output_tokens":2723233,"total_runtime_s":92839.4148778622,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.90574125005675,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bf0a3d117f3e6c569484","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ecd93dcacb1b60651f578cf9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21559999999999999,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":624,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[311];split=licensed;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":624,"metadata":{"avgScore":0.21559999999999999,"avg_score":0.21559999999999999,"avg_score_percent":21.56,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":208.55,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":320986154,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":479276,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":5,"runtime":26915.112756414877,"source_flags":[],"source_harness":"ale_claw","source_row_index":311,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":208.55,"totalDurationS":26915.112756414877,"totalInputTokens":320986154,"totalOutputTokens":479276,"total_cost_usd":208.55,"total_input_tokens":320986154,"total_output_tokens":479276,"total_runtime_s":26915.112756414877,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.56,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c4455914ae3e31f289e6","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ea028351bca04bf8b06c7460","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3521506822590183,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":84,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[41];split=full;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":84,"metadata":{"avgScore":0.3521506822590183,"avg_score":0.3521506822590183,"avg_score_percent":35.21506822590183,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1688.85,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":807329919,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3963995.5,"passRate":0.15789473684210525,"pass_rate":0.15789473684210525,"pass_rate_percent":15.789473684210526,"passes":24,"runs":142,"runtime":446984.7452694551,"source_flags":[],"source_harness":"openclaw","source_row_index":41,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1688.85,"totalDurationS":446984.7452694551,"totalInputTokens":807329919,"totalOutputTokens":3963995.5,"total_cost_usd":1688.85,"total_input_tokens":807329919,"total_output_tokens":3963995.5,"total_runtime_s":446984.7452694551,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.21506822590183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0f6f7765c918b57ef82c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e863798b8d422d3bba998e76","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3469339382470838,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":796,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[397];split=linux_only;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":796,"metadata":{"avgScore":0.3469339382470838,"avg_score":0.3469339382470838,"avg_score_percent":34.69339382470838,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":709.97,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":176866012,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1651850,"passRate":0.1523809523809524,"pass_rate":0.1523809523809524,"pass_rate_percent":15.238095238095239,"passes":17,"runs":74,"runtime":69448.19969177898,"source_flags":[],"source_harness":"droid","source_row_index":397,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":70,"tokens":null,"totalCostUsd":709.97,"totalDurationS":69448.19969177898,"totalInputTokens":176866012,"totalOutputTokens":1651850,"total_cost_usd":709.97,"total_input_tokens":176866012,"total_output_tokens":1651850,"total_runtime_s":69448.19969177898,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":34.69339382470838,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-abcf129d15826bef5a5e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e50f7229dcb59148e3d08337","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2777319707686421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1038,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[518];split=unlicensed/full-spectrum;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1038,"metadata":{"avgScore":0.2777319707686421,"avg_score":0.2777319707686421,"avg_score_percent":27.773197076864207,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":301.67,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":126991014,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1028842,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":5,"runs":48,"runtime":81810.31944376754,"source_flags":[],"source_harness":"openclaw","source_row_index":518,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":301.67,"totalDurationS":81810.31944376754,"totalInputTokens":126991014,"totalOutputTokens":1028842,"total_cost_usd":301.67,"total_input_tokens":126991014,"total_output_tokens":1028842,"total_runtime_s":81810.31944376754,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":27.773197076864207,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1af9edce1b5cb80811d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e0dd9e3f41096f92d0ba963d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6045069241062173,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":420,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[209];split=full/near-term;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":420,"metadata":{"avgScore":0.6045069241062173,"avg_score":0.6045069241062173,"avg_score_percent":60.450692410621734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":311.72,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":359152755,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1944556,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":74416.76165142255,"source_flags":[],"source_harness":"ale_claw","source_row_index":209,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":311.72,"totalDurationS":74416.76165142255,"totalInputTokens":359152755,"totalOutputTokens":1944556,"total_cost_usd":311.72,"total_input_tokens":359152755,"total_output_tokens":1944556,"total_runtime_s":74416.76165142255,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":60.450692410621734,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5d2547deaff5c94966ba","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dfa925c0bb276ef97190a980","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08029413579144805,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":316,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[157];split=full/last-exam;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":316,"metadata":{"avgScore":0.08029413579144805,"avg_score":0.08029413579144805,"avg_score_percent":8.029413579144805,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":495.79,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":145716643,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":920145,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":20,"runtime":37053.47873265087,"source_flags":[],"source_harness":"droid","source_row_index":157,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":20,"tokens":null,"totalCostUsd":495.79,"totalDurationS":37053.47873265087,"totalInputTokens":145716643,"totalOutputTokens":920145,"total_cost_usd":495.79,"total_input_tokens":145716643,"total_output_tokens":920145,"total_runtime_s":37053.47873265087,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.029413579144805,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-11d810e629303d420fd7","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ca3b8bb7c49d9ebefeb09a0a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.42487183710893694,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":48,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[23];split=full;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":48,"metadata":{"avgScore":0.42487183710893694,"avg_score":0.42487183710893694,"avg_score_percent":42.487183710893696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1899.32,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":429515271,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":4464514,"passRate":0.21052631578947367,"pass_rate":0.21052631578947367,"pass_rate_percent":21.052631578947366,"passes":32,"runs":137,"runtime":229732.552303883,"source_flags":[],"source_harness":"cursor_cli","source_row_index":23,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1899.32,"totalDurationS":229732.552303883,"totalInputTokens":429515271,"totalOutputTokens":4464514,"total_cost_usd":1899.32,"total_input_tokens":429515271,"total_output_tokens":4464514,"total_runtime_s":229732.552303883,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.487183710893696,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-998a3dc922b74b52cdd3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c29049eaeb53d0fb883b4c93","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.10029435149513931,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":302,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[150];split=full/last-exam;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":302,"metadata":{"avgScore":0.10029435149513931,"avg_score":0.10029435149513931,"avg_score_percent":10.029435149513931,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":684.93,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":170082318,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1387243,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":61575.264092724305,"source_flags":[],"source_harness":"claude_code","source_row_index":150,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":684.93,"totalDurationS":61575.264092724305,"totalInputTokens":170082318,"totalOutputTokens":1387243,"total_cost_usd":684.93,"total_input_tokens":170082318,"total_output_tokens":1387243,"total_runtime_s":61575.264092724305,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.029435149513931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f5951d369903b51d8ffd","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c179e8c659923f013b52d9a0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09501958811261163,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1068,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[533];split=unlicensed/full-spectrum;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1068,"metadata":{"avgScore":0.09501958811261163,"avg_score":0.09501958811261163,"avg_score_percent":9.501958811261163,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":150.16,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":36489552,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":423717,"passRate":0.02,"pass_rate":0.02,"pass_rate_percent":2.0,"passes":1,"runs":18,"runtime":10051.606426970102,"source_flags":[],"source_harness":"droid","source_row_index":533,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":18,"tokens":null,"totalCostUsd":150.16,"totalDurationS":10051.606426970102,"totalInputTokens":36489552,"totalOutputTokens":423717,"total_cost_usd":150.16,"total_input_tokens":36489552,"total_output_tokens":423717,"total_runtime_s":10051.606426970102,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.501958811261163,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f480408f8581c39f82b5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bd57dc9d8252b7479add01a9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.44713736097477313,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":880,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[439];split=unlicensed;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":880,"metadata":{"avgScore":0.44713736097477313,"avg_score":0.44713736097477313,"avg_score_percent":44.713736097477316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1564.2,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":317070704,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3915017,"passRate":0.22695035460992907,"pass_rate":0.22695035460992907,"pass_rate_percent":22.69503546099291,"passes":32,"runs":130,"runtime":209777.7502665911,"source_flags":[],"source_harness":"cursor_cli","source_row_index":439,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1564.2,"totalDurationS":209777.7502665911,"totalInputTokens":317070704,"totalOutputTokens":3915017,"total_cost_usd":1564.2,"total_input_tokens":317070704,"total_output_tokens":3915017,"total_runtime_s":209777.7502665911,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":44.713736097477316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-62f9aab0d0e95a04eba5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b776397b72e4e7de4ebc3b25","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4326975564273962,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":768,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[383];split=linux_only;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":768,"metadata":{"avgScore":0.4326975564273962,"avg_score":0.4326975564273962,"avg_score_percent":43.26975564273962,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.55,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":389612585,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2906411,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":21,"runs":103,"runtime":159298.15374917473,"source_flags":[],"source_harness":"ale_claw","source_row_index":383,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":415.55,"totalDurationS":159298.15374917473,"totalInputTokens":389612585,"totalOutputTokens":2906411,"total_cost_usd":415.55,"total_input_tokens":389612585,"total_output_tokens":2906411,"total_runtime_s":159298.15374917473,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":43.26975564273962,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-90e920657d7a0f4f562f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b401b7d0e8ec93bc5927225a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.33663835749551463,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1398,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[698];split=unlicensed/overall;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1398,"metadata":{"avgScore":0.33663835749551463,"avg_score":0.33663835749551463,"avg_score_percent":33.66383574955146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1325.06,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":344833331.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2642831,"passRate":0.1453900709219858,"pass_rate":0.1453900709219858,"pass_rate_percent":14.539007092198581,"passes":22,"runs":101,"runtime":99274.81146285753,"source_flags":[],"source_harness":"droid","source_row_index":698,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":95,"tokens":null,"totalCostUsd":1325.06,"totalDurationS":99274.81146285753,"totalInputTokens":344833331.5,"totalOutputTokens":2642831,"total_cost_usd":1325.06,"total_input_tokens":344833331.5,"total_output_tokens":2642831,"total_runtime_s":99274.81146285753,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.66383574955146,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ca8a5643c00f8df80746","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ab2bc7ea59d5870c5d778cc9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3745618668551661,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1390,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[694];split=unlicensed/overall;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1390,"metadata":{"avgScore":0.3745618668551661,"avg_score":0.3745618668551661,"avg_score_percent":37.45618668551661,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1495.3,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":722150808,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3808451.5,"passRate":0.1702127659574468,"pass_rate":0.1702127659574468,"pass_rate_percent":17.02127659574468,"passes":24,"runs":137,"runtime":403839.59668815427,"source_flags":[],"source_harness":"openclaw","source_row_index":694,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1495.3,"totalDurationS":403839.59668815427,"totalInputTokens":722150808,"totalOutputTokens":3808451.5,"total_cost_usd":1495.3,"total_input_tokens":722150808,"total_output_tokens":3808451.5,"total_runtime_s":403839.59668815427,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.45618668551661,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-934447a9152f9e74e93d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a0b111b7ff22e1b8c8b11bd5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.35803700118782017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":82,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[40];split=full;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":82,"metadata":{"avgScore":0.35803700118782017,"avg_score":0.35803700118782017,"avg_score_percent":35.80370011878202,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1792.72,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":456438963,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3700468,"passRate":0.13815789473684212,"pass_rate":0.13815789473684212,"pass_rate_percent":13.815789473684212,"passes":21,"runs":137,"runtime":153370.83306965837,"source_flags":[],"source_harness":"claude_code","source_row_index":40,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1792.72,"totalDurationS":153370.83306965837,"totalInputTokens":456438963,"totalOutputTokens":3700468,"total_cost_usd":1792.72,"total_input_tokens":456438963,"total_output_tokens":3700468,"total_runtime_s":153370.83306965837,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.80370011878202,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1e0e709c458e4097b0f3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9ecdd6f20ba3da71c088478d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4049605336355299,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":536,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[267];split=full/overall;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":536,"metadata":{"avgScore":0.4049605336355299,"avg_score":0.4049605336355299,"avg_score_percent":40.49605336355299,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1132.24,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1342066095,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":5568247,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":143,"runtime":279527.28043104644,"source_flags":[],"source_harness":"ale_claw","source_row_index":267,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1132.24,"totalDurationS":279527.28043104644,"totalInputTokens":1342066095,"totalOutputTokens":5568247,"total_cost_usd":1132.24,"total_input_tokens":1342066095,"total_output_tokens":5568247,"total_runtime_s":279527.28043104644,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":40.49605336355299,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8c62ae54f6eb20f98968","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9cbfdcdeb8f995dbecd05548","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5799257416141242,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":434,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[216];split=full/near-term;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":434,"metadata":{"avgScore":0.5799257416141242,"avg_score":0.5799257416141242,"avg_score_percent":57.99257416141243,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":508.37,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":201660470,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1516121.5,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":66,"runtime":171722.06837378268,"source_flags":[],"source_harness":"openclaw","source_row_index":216,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":508.37,"totalDurationS":171722.06837378268,"totalInputTokens":201660470,"totalOutputTokens":1516121.5,"total_cost_usd":508.37,"total_input_tokens":201660470,"total_output_tokens":1516121.5,"total_runtime_s":171722.06837378268,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":57.99257416141243,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8d291dabcff76c8dc10","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-97ad07831e1d6afb344169dc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06468889570193674,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":682,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[340];split=licensed;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":682,"metadata":{"avgScore":0.06468889570193674,"avg_score":0.06468889570193674,"avg_score_percent":6.468889570193674,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.19,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":7226933,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":83461,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":4,"runtime":2505.957411190262,"source_flags":[],"source_harness":"droid","source_row_index":340,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":4,"tokens":null,"totalCostUsd":31.19,"totalDurationS":2505.957411190262,"totalInputTokens":7226933,"totalOutputTokens":83461,"total_cost_usd":31.19,"total_input_tokens":7226933,"total_output_tokens":83461,"total_runtime_s":2505.957411190262,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.468889570193674,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c64a86d01d4cf2ec920f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-96f79550e286c84814832685","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.36513474394967993,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1012,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[505];split=unlicensed/full-spectrum;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1012,"metadata":{"avgScore":0.36513474394967993,"avg_score":0.36513474394967993,"avg_score_percent":36.51347439496799,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":201.89,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":222692155,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1493759,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":49,"runtime":47608.04235754142,"source_flags":[],"source_harness":"ale_claw","source_row_index":505,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":201.89,"totalDurationS":47608.04235754142,"totalInputTokens":222692155,"totalOutputTokens":1493759,"total_cost_usd":201.89,"total_input_tokens":222692155,"total_output_tokens":1493759,"total_runtime_s":47608.04235754142,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":36.51347439496799,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6e924d7c9e3b9a5bdfeb","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8fd38db763813fd8fd9c4cec","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.45680671377661547,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":752,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[375];split=linux_only;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":752,"metadata":{"avgScore":0.45680671377661547,"avg_score":0.45680671377661547,"avg_score_percent":45.68067137766155,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1044.57,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":207544892,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2506674,"passRate":0.22857142857142856,"pass_rate":0.22857142857142856,"pass_rate_percent":22.857142857142858,"passes":24,"runs":96,"runtime":89632.5721885981,"source_flags":[],"source_harness":"cursor_cli","source_row_index":375,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":96,"tokens":null,"totalCostUsd":1044.57,"totalDurationS":89632.5721885981,"totalInputTokens":207544892,"totalOutputTokens":2506674,"total_cost_usd":1044.57,"total_input_tokens":207544892,"total_output_tokens":2506674,"total_runtime_s":89632.5721885981,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.68067137766155,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fcb644906e550d2b86c5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-847abf3e963dffb151adc68e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.39292900253171503,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":996,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[497];split=unlicensed/full-spectrum;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":996,"metadata":{"avgScore":0.39292900253171503,"avg_score":0.39292900253171503,"avg_score_percent":39.2929002531715,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":580.62,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":128052474,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1427847,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":48,"runtime":40744.41926420876,"source_flags":[],"source_harness":"cursor_cli","source_row_index":497,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":580.62,"totalDurationS":40744.41926420876,"totalInputTokens":128052474,"totalOutputTokens":1427847,"total_cost_usd":580.62,"total_input_tokens":128052474,"total_output_tokens":1427847,"total_runtime_s":40744.41926420876,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":39.2929002531715,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4581f1a694e0832f2782","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-81464099b2aa310158c7a509","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3745618668551661,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":918,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[458];split=unlicensed;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":918,"metadata":{"avgScore":0.3745618668551661,"avg_score":0.3745618668551661,"avg_score_percent":37.45618668551661,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1495.3,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":722150808,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3808451.5,"passRate":0.1702127659574468,"pass_rate":0.1702127659574468,"pass_rate_percent":17.02127659574468,"passes":24,"runs":137,"runtime":403839.59668815427,"source_flags":[],"source_harness":"openclaw","source_row_index":458,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1495.3,"totalDurationS":403839.59668815427,"totalInputTokens":722150808,"totalOutputTokens":3808451.5,"total_cost_usd":1495.3,"total_input_tokens":722150808,"total_output_tokens":3808451.5,"total_runtime_s":403839.59668815427,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.45618668551661,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-56762525f66523bd4a25","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7eb6bcaee41817435e338939","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.274703254822263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":204,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[101];split=full/full-spectrum;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":204,"metadata":{"avgScore":0.274703254822263,"avg_score":0.274703254822263,"avg_score_percent":27.4703254822263,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":369.56,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":162969004,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1135215,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":51,"runtime":107204.82072244062,"source_flags":[],"source_harness":"openclaw","source_row_index":101,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":51,"tokens":null,"totalCostUsd":369.56,"totalDurationS":107204.82072244062,"totalInputTokens":162969004,"totalOutputTokens":1135215,"total_cost_usd":369.56,"total_input_tokens":162969004,"total_output_tokens":1135215,"total_runtime_s":107204.82072244062,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":27.4703254822263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6d2211e266e45f87eb43","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7e103d6a00daa5e55d001378","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.043412991868513016,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":334,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[166];split=full/last-exam;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":334,"metadata":{"avgScore":0.043412991868513016,"avg_score":0.043412991868513016,"avg_score_percent":4.341299186851302,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":841.65,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":452898006,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1462721,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":32,"runtime":180025.200925768,"source_flags":[],"source_harness":"openclaw","source_row_index":166,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":841.65,"totalDurationS":180025.200925768,"totalInputTokens":452898006,"totalOutputTokens":1462721,"total_cost_usd":841.65,"total_input_tokens":452898006,"total_output_tokens":1462721,"total_runtime_s":180025.200925768,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.341299186851302,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1b0a3b168f442d3d16dc","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-74898e0281548fd520c25d6b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.3722051359473687,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1392,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[695];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1392,"metadata":{"avgScore":0.3722051359473687,"avg_score":0.3722051359473687,"avg_score_percent":37.22051359473687,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1558.83,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":380178747,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3444699,"passRate":0.14893617021276595,"pass_rate":0.14893617021276595,"pass_rate_percent":14.893617021276595,"passes":21,"runs":130,"runtime":141696.4445367267,"source_flags":[],"source_harness":"claude_code","source_row_index":695,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1558.83,"totalDurationS":141696.4445367267,"totalInputTokens":380178747,"totalOutputTokens":3444699,"total_cost_usd":1558.83,"total_input_tokens":380178747,"total_output_tokens":3444699,"total_runtime_s":141696.4445367267,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.22051359473687,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-481f62b2d265e21385d4","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-71c358349b11f106a74c61b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.31695780433940046,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":562,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[280];split=full/overall;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":562,"metadata":{"avgScore":0.31695780433940046,"avg_score":0.31695780433940046,"avg_score_percent":31.695780433940048,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1356.25,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":352060264.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2726292,"passRate":0.13486842105263158,"pass_rate":0.13486842105263158,"pass_rate_percent":13.486842105263158,"passes":22,"runs":105,"runtime":101780.76887404779,"source_flags":[],"source_harness":"droid","source_row_index":280,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":99,"tokens":null,"totalCostUsd":1356.25,"totalDurationS":101780.76887404779,"totalInputTokens":352060264.5,"totalOutputTokens":2726292,"total_cost_usd":1356.25,"total_input_tokens":352060264.5,"total_output_tokens":2726292,"total_runtime_s":101780.76887404779,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.695780433940048,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0029cef5d882add847e3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6e2a04a5c14e463fce17fe6b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08531923866916752,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1152,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[575];split=unlicensed/last-exam;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1152,"metadata":{"avgScore":0.08531923866916752,"avg_score":0.08531923866916752,"avg_score_percent":8.531923866916753,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":475.55,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":141000329,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":865873,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":17,"runtime":35455.30997874378,"source_flags":[],"source_harness":"droid","source_row_index":575,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":17,"tokens":null,"totalCostUsd":475.55,"totalDurationS":35455.30997874378,"totalInputTokens":141000329,"totalOutputTokens":865873,"total_cost_usd":475.55,"total_input_tokens":141000329,"total_output_tokens":865873,"total_runtime_s":35455.30997874378,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.531923866916753,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c0bf8283dded230f385c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6c11847911d9ccb96d782921","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12847049593392926,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1110,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[554];split=unlicensed/last-exam;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1110,"metadata":{"avgScore":0.12847049593392926,"avg_score":0.12847049593392926,"avg_score_percent":12.847049593392926,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":496.57,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":92176971,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1046938,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":22,"runtime":58179.24658737797,"source_flags":[],"source_harness":"cursor_cli","source_row_index":554,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":22,"tokens":null,"totalCostUsd":496.57,"totalDurationS":58179.24658737797,"totalInputTokens":92176971,"totalOutputTokens":1046938,"total_cost_usd":496.57,"total_input_tokens":92176971,"total_output_tokens":1046938,"total_runtime_s":58179.24658737797,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.847049593392926,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8584677fea68912b403d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-64f887844a4c590ebc6d3870","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.44713736097477313,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1352,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[675];split=unlicensed/overall;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1352,"metadata":{"avgScore":0.44713736097477313,"avg_score":0.44713736097477313,"avg_score_percent":44.713736097477316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1564.2,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":317070704,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3915017,"passRate":0.22695035460992907,"pass_rate":0.22695035460992907,"pass_rate_percent":22.69503546099291,"passes":32,"runs":130,"runtime":209777.7502665911,"source_flags":[],"source_harness":"cursor_cli","source_row_index":675,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1564.2,"totalDurationS":209777.7502665911,"totalInputTokens":317070704,"totalOutputTokens":3915017,"total_cost_usd":1564.2,"total_input_tokens":317070704,"total_output_tokens":3915017,"total_runtime_s":209777.7502665911,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":44.713736097477316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1fff72fc366e38018095","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5b120107a614619d2581a6a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.17642727381542536,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":642,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[320];split=licensed;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":642,"metadata":{"avgScore":0.17642727381542536,"avg_score":0.17642727381542536,"avg_score_percent":17.642727381542535,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":233.89,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":76260216,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":255769,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":11674.388532931684,"source_flags":[],"source_harness":"claude_code","source_row_index":320,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":233.89,"totalDurationS":11674.388532931684,"totalInputTokens":76260216,"totalOutputTokens":255769,"total_cost_usd":233.89,"total_input_tokens":76260216,"total_output_tokens":255769,"total_runtime_s":11674.388532931684,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.642727381542535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-02e04e1e5b0d726b809f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-51f6999d97f8d80ca0eab429","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.07863300882269293,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":318,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[158];split=full/last-exam;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":318,"metadata":{"avgScore":0.07863300882269293,"avg_score":0.07863300882269293,"avg_score_percent":7.863300882269293,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":593.61,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":707313906,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2192916,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":155269.84613300068,"source_flags":[],"source_harness":"ale_claw","source_row_index":158,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":593.61,"totalDurationS":155269.84613300068,"totalInputTokens":707313906,"totalOutputTokens":2192916,"total_cost_usd":593.61,"total_input_tokens":707313906,"total_output_tokens":2192916,"total_runtime_s":155269.84613300068,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":7.863300882269293,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c5489022014fe98376e1","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-516c2a107b925598f135e53b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.29108336390143064,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":198,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[98];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":198,"metadata":{"avgScore":0.29108336390143064,"avg_score":0.29108336390143064,"avg_score_percent":29.108336390143062,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":747.4,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":201998860,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1546037,"passRate":0.12727272727272726,"pass_rate":0.12727272727272726,"pass_rate_percent":12.727272727272727,"passes":7,"runs":50,"runtime":43471.418552602874,"source_flags":[],"source_harness":"claude_code","source_row_index":98,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":747.4,"totalDurationS":43471.418552602874,"totalInputTokens":201998860,"totalOutputTokens":1546037,"total_cost_usd":747.4,"total_input_tokens":201998860,"total_output_tokens":1546037,"total_runtime_s":43471.418552602874,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.108336390143062,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1b3b8ad1b13399aeed5b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4d593e6fdfae609ee5c90d05","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.27546770029157375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1040,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[519];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1040,"metadata":{"avgScore":0.27546770029157375,"avg_score":0.27546770029157375,"avg_score_percent":27.546770029157376,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":559.29,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":138650190,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1350638,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":46,"runtime":34153.082216965966,"source_flags":[],"source_harness":"claude_code","source_row_index":519,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":46,"tokens":null,"totalCostUsd":559.29,"totalDurationS":34153.082216965966,"totalInputTokens":138650190,"totalOutputTokens":1350638,"total_cost_usd":559.29,"total_input_tokens":138650190,"total_output_tokens":1350638,"total_runtime_s":34153.082216965966,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.546770029157376,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7a8657c06c45437721f4","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4aaa5b5a62667f60b8e3bc98","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6269364006481253,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1242,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[620];split=unlicensed/near-term;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1242,"metadata":{"avgScore":0.6269364006481253,"avg_score":0.6269364006481253,"avg_score_percent":62.69364006481253,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":557.68,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":110261629,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1648119,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":120381.51676356932,"source_flags":[],"source_harness":"cursor_cli","source_row_index":620,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":557.68,"totalDurationS":120381.51676356932,"totalInputTokens":110261629,"totalOutputTokens":1648119,"total_cost_usd":557.68,"total_input_tokens":110261629,"total_output_tokens":1648119,"total_runtime_s":120381.51676356932,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":62.69364006481253,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c502fdb57aa039d77d3d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-46063b716c35c7adac808c00","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06488004334475987,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":678,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[338];split=licensed;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":678,"metadata":{"avgScore":0.06488004334475987,"avg_score":0.06488004334475987,"avg_score_percent":6.4880043344759875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":193.55,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":85179111,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":155544,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":43145.14858130082,"source_flags":[],"source_harness":"openclaw","source_row_index":338,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":193.55,"totalDurationS":43145.14858130082,"totalInputTokens":85179111,"totalOutputTokens":155544,"total_cost_usd":193.55,"total_input_tokens":85179111,"total_output_tokens":155544,"total_runtime_s":43145.14858130082,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.4880043344759875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-72b59d5d112556024ae6","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4557038fef489e581a97f46e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6171275522572685,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":414,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[206];split=full/near-term;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":414,"metadata":{"avgScore":0.6171275522572685,"avg_score":0.6171275522572685,"avg_score_percent":61.712755225726845,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":737.55,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":176244296.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1470123,"passRate":0.291044776119403,"pass_rate":0.291044776119403,"pass_rate_percent":29.1044776119403,"passes":21,"runs":69,"runtime":57944.460609696805,"source_flags":[],"source_harness":"droid","source_row_index":206,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":737.55,"totalDurationS":57944.460609696805,"totalInputTokens":176244296.5,"totalOutputTokens":1470123,"total_cost_usd":737.55,"total_input_tokens":176244296.5,"total_output_tokens":1470123,"total_runtime_s":57944.460609696805,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":61.712755225726845,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-51536293063904d1d5de","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4425913ddbe87b1883f701f0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.33663835749551463,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":926,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[462];split=unlicensed;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":926,"metadata":{"avgScore":0.33663835749551463,"avg_score":0.33663835749551463,"avg_score_percent":33.66383574955146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1325.06,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":344833331.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2642831,"passRate":0.1453900709219858,"pass_rate":0.1453900709219858,"pass_rate_percent":14.539007092198581,"passes":22,"runs":101,"runtime":99274.81146285753,"source_flags":[],"source_harness":"droid","source_row_index":462,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":95,"tokens":null,"totalCostUsd":1325.06,"totalDurationS":99274.81146285753,"totalInputTokens":344833331.5,"totalOutputTokens":2642831,"total_cost_usd":1325.06,"total_input_tokens":344833331.5,"total_output_tokens":2642831,"total_runtime_s":99274.81146285753,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.66383574955146,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4c639543ca33fab2cd8e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-421f4c4e093a1542a9a4a284","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11436091784257574,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":282,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[140];split=full/last-exam;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":282,"metadata":{"avgScore":0.11436091784257574,"avg_score":0.11436091784257574,"avg_score_percent":11.436091784257574,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":587.73,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":121011885,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1184118,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":26,"runtime":66683.20773970685,"source_flags":[],"source_harness":"cursor_cli","source_row_index":140,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":26,"tokens":null,"totalCostUsd":587.73,"totalDurationS":66683.20773970685,"totalInputTokens":121011885,"totalOutputTokens":1184118,"total_cost_usd":587.73,"total_input_tokens":121011885,"total_output_tokens":1184118,"total_runtime_s":66683.20773970685,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.436091784257574,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cf294ceba859edeeecd3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3afdafbe3b5b3f1f478adbf7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.35803700118782017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":554,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[276];split=full/overall;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":554,"metadata":{"avgScore":0.35803700118782017,"avg_score":0.35803700118782017,"avg_score_percent":35.80370011878202,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1792.72,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":456438963,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3700468,"passRate":0.13815789473684212,"pass_rate":0.13815789473684212,"pass_rate_percent":13.815789473684212,"passes":21,"runs":137,"runtime":153370.83306965837,"source_flags":[],"source_harness":"claude_code","source_row_index":276,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1792.72,"totalDurationS":153370.83306965837,"totalInputTokens":456438963,"totalOutputTokens":3700468,"total_cost_usd":1792.72,"total_input_tokens":456438963,"total_output_tokens":3700468,"total_runtime_s":153370.83306965837,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.80370011878202,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8b0a98cb7773bc96895e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-394fe4569e8c9e6e2c204d7e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.38739364141550475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":792,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[395];split=linux_only;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":792,"metadata":{"avgScore":0.38739364141550475,"avg_score":0.38739364141550475,"avg_score_percent":38.739364141550475,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":959.57,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":516245034,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2766369.5,"passRate":0.17142857142857143,"pass_rate":0.17142857142857143,"pass_rate_percent":17.142857142857142,"passes":18,"runs":104,"runtime":192542.93685384613,"source_flags":[],"source_harness":"openclaw","source_row_index":395,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":959.57,"totalDurationS":192542.93685384613,"totalInputTokens":516245034,"totalOutputTokens":2766369.5,"total_cost_usd":959.57,"total_input_tokens":516245034,"total_output_tokens":2766369.5,"total_runtime_s":192542.93685384613,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":38.739364141550475,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-34014f21a3f016448707","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-37c3301ccc9c2ddf82382c9f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03860312303344982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1174,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[586];split=unlicensed/last-exam;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1174,"metadata":{"avgScore":0.03860312303344982,"avg_score":0.03860312303344982,"avg_score_percent":3.8603123033449824,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":714.1,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":403184810,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1398987,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":29,"runtime":160106.59431680728,"source_flags":[],"source_harness":"openclaw","source_row_index":586,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":714.1,"totalDurationS":160106.59431680728,"totalInputTokens":403184810,"totalOutputTokens":1398987,"total_cost_usd":714.1,"total_input_tokens":403184810,"total_output_tokens":1398987,"total_runtime_s":160106.59431680728,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.8603123033449824,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cc47788eec73d9668106","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-30f90eb4047cd09e94c283cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6269364006481253,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":406,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[202];split=full/near-term;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":406,"metadata":{"avgScore":0.6269364006481253,"avg_score":0.6269364006481253,"avg_score_percent":62.69364006481253,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":557.68,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":110261629,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1648119,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":120381.51676356932,"source_flags":[],"source_harness":"cursor_cli","source_row_index":202,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":557.68,"totalDurationS":120381.51676356932,"totalInputTokens":110261629,"totalOutputTokens":1648119,"total_cost_usd":557.68,"total_input_tokens":110261629,"total_output_tokens":1648119,"total_runtime_s":120381.51676356932,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":62.69364006481253,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e6b5b66c8953304d6a3e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2f358d40b96b1f2859a6df8c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.42487183710893694,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":520,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[259];split=full/overall;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":520,"metadata":{"avgScore":0.42487183710893694,"avg_score":0.42487183710893694,"avg_score_percent":42.487183710893696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1899.32,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":429515271,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":4464514,"passRate":0.21052631578947367,"pass_rate":0.21052631578947367,"pass_rate_percent":21.052631578947366,"passes":32,"runs":137,"runtime":229732.552303883,"source_flags":[],"source_harness":"cursor_cli","source_row_index":259,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1899.32,"totalDurationS":229732.552303883,"totalInputTokens":429515271,"totalOutputTokens":4464514,"total_cost_usd":1899.32,"total_input_tokens":429515271,"total_output_tokens":4464514,"total_runtime_s":229732.552303883,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.487183710893696,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d0552b65aee006720e59","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2732077b4b795ea20ff74d50","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5799257416141242,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1270,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[634];split=unlicensed/near-term;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1270,"metadata":{"avgScore":0.5799257416141242,"avg_score":0.5799257416141242,"avg_score_percent":57.99257416141243,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":508.37,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":201660470,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1516121.5,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":66,"runtime":171722.06837378268,"source_flags":[],"source_harness":"openclaw","source_row_index":634,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":508.37,"totalDurationS":171722.06837378268,"totalInputTokens":201660470,"totalOutputTokens":1516121.5,"total_cost_usd":508.37,"total_input_tokens":201660470,"total_output_tokens":1516121.5,"total_runtime_s":171722.06837378268,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":57.99257416141243,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-23e8792a0bf9aa1e2b10","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1e0cecc9706c28ddf11fb728","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.08768284482818715,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1148,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[573];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1148,"metadata":{"avgScore":0.08768284482818715,"avg_score":0.08768284482818715,"avg_score_percent":8.768284482818716,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":576.53,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":134441437,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1284319,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":26,"runtime":57084.972001090646,"source_flags":[],"source_harness":"claude_code","source_row_index":573,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":576.53,"totalDurationS":57084.972001090646,"totalInputTokens":134441437,"totalOutputTokens":1284319,"total_cost_usd":576.53,"total_input_tokens":134441437,"total_output_tokens":1284319,"total_runtime_s":57084.972001090646,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":8.768284482818716,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-72a75731fba2e387f7ed","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1b7e78f663cbd5b748710d50","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08321514450874437,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1154,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[576];split=unlicensed/last-exam;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1154,"metadata":{"avgScore":0.08321514450874437,"avg_score":0.08321514450874437,"avg_score_percent":8.321514450874437,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":425.56,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":450451674,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1846704,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":28,"runtime":135702.5551899186,"source_flags":[],"source_harness":"ale_claw","source_row_index":576,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":425.56,"totalDurationS":135702.5551899186,"totalInputTokens":450451674,"totalOutputTokens":1846704,"total_cost_usd":425.56,"total_input_tokens":450451674,"total_output_tokens":1846704,"total_runtime_s":135702.5551899186,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.321514450874437,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a3942ed9ddc42cf95c3c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-174845d1e43255181cd17438","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.37616366308547566,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":168,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[83];split=full/full-spectrum;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":168,"metadata":{"avgScore":0.37616366308547566,"avg_score":0.37616366308547566,"avg_score_percent":37.61636630854757,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":841.57,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":215576200,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1889131,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":53,"runtime":53276.67023467668,"source_flags":[],"source_harness":"cursor_cli","source_row_index":83,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":841.57,"totalDurationS":53276.67023467668,"totalInputTokens":215576200,"totalOutputTokens":1889131,"total_cost_usd":841.57,"total_input_tokens":215576200,"total_output_tokens":1889131,"total_runtime_s":53276.67023467668,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.61636630854757,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6705ecd837c7abb35cfe","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-15b865aa0617fe67148130d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.31695780433940046,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":90,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[44];split=full;harness=droid;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":90,"metadata":{"avgScore":0.31695780433940046,"avg_score":0.31695780433940046,"avg_score_percent":31.695780433940048,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1356.25,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":352060264.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":2726292,"passRate":0.13486842105263158,"pass_rate":0.13486842105263158,"pass_rate_percent":13.486842105263158,"passes":22,"runs":105,"runtime":101780.76887404779,"source_flags":[],"source_harness":"droid","source_row_index":44,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":99,"tokens":null,"totalCostUsd":1356.25,"totalDurationS":101780.76887404779,"totalInputTokens":352060264.5,"totalOutputTokens":2726292,"total_cost_usd":1356.25,"total_input_tokens":352060264.5,"total_output_tokens":2726292,"total_runtime_s":101780.76887404779,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.695780433940048,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7378d0347421e43cb09a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1569ed04586695c65b80b489","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3661224944997091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":174,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[86];split=full/full-spectrum;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":174,"metadata":{"avgScore":0.3661224944997091,"avg_score":0.3661224944997091,"avg_score_percent":36.61224944997091,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.46,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":288810311,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1669212,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":52,"runtime":55840.03182199924,"source_flags":[],"source_harness":"ale_claw","source_row_index":86,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":52,"tokens":null,"totalCostUsd":246.46,"totalDurationS":55840.03182199924,"totalInputTokens":288810311,"totalOutputTokens":1669212,"total_cost_usd":246.46,"total_input_tokens":288810311,"total_output_tokens":1669212,"total_runtime_s":55840.03182199924,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":36.61224944997091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fffad1c505869e21151f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-143c853fc50c916a46ed203c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.3722051359473687,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":920,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[459];split=unlicensed;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":920,"metadata":{"avgScore":0.3722051359473687,"avg_score":0.3722051359473687,"avg_score_percent":37.22051359473687,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1558.83,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":380178747,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3444699,"passRate":0.14893617021276595,"pass_rate":0.14893617021276595,"pass_rate_percent":14.893617021276595,"passes":21,"runs":130,"runtime":141696.4445367267,"source_flags":[],"source_harness":"claude_code","source_row_index":459,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1558.83,"totalDurationS":141696.4445367267,"totalInputTokens":380178747,"totalOutputTokens":3444699,"total_cost_usd":1558.83,"total_input_tokens":380178747,"total_output_tokens":3444699,"total_runtime_s":141696.4445367267,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.22051359473687,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-31014f8ecc9a9c38c37d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1363bdc286239d81754ab36b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.5578444176897955,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":438,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[218];split=full/near-term;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":438,"metadata":{"avgScore":0.5578444176897955,"avg_score":0.5578444176897955,"avg_score_percent":55.784441768979555,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":496.33,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":120384368,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1068334,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":63,"runtime":58636.88796496298,"source_flags":[],"source_harness":"claude_code","source_row_index":218,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":496.33,"totalDurationS":58636.88796496298,"totalInputTokens":120384368,"totalOutputTokens":1068334,"total_cost_usd":496.33,"total_input_tokens":120384368,"total_output_tokens":1068334,"total_runtime_s":58636.88796496298,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":55.784441768979555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fcf75aa17c63641dd971","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-127e9abde86dd3ee9a7608a1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.5578444176897955,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1274,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[636];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1274,"metadata":{"avgScore":0.5578444176897955,"avg_score":0.5578444176897955,"avg_score_percent":55.784441768979555,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":496.33,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":120384368,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":1068334,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":63,"runtime":58636.88796496298,"source_flags":[],"source_harness":"claude_code","source_row_index":636,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":496.33,"totalDurationS":58636.88796496298,"totalInputTokens":120384368,"totalOutputTokens":1068334,"total_cost_usd":496.33,"total_input_tokens":120384368,"total_output_tokens":1068334,"total_runtime_s":58636.88796496298,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":55.784441768979555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-921744aa856173254308","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-10601c5ee87f6b55327bf29a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13946830391958281,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":652,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[325];split=licensed;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":652,"metadata":{"avgScore":0.13946830391958281,"avg_score":0.13946830391958281,"avg_score_percent":13.946830391958281,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":335.13,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":112444567,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":549497,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":19954.802037291927,"source_flags":[],"source_harness":"cursor_cli","source_row_index":325,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":335.13,"totalDurationS":19954.802037291927,"totalInputTokens":112444567,"totalOutputTokens":549497,"total_cost_usd":335.13,"total_input_tokens":112444567,"total_output_tokens":549497,"total_runtime_s":19954.802037291927,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.946830391958281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-971e3063f98ee085fa3a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0fe65268bde9c56bd939b67d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4197333412241174,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1376,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[687];split=unlicensed/overall;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1376,"metadata":{"avgScore":0.4197333412241174,"avg_score":0.4197333412241174,"avg_score_percent":41.973334122411735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":923.69,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1021079941,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":5088971,"passRate":0.19148936170212766,"pass_rate":0.19148936170212766,"pass_rate_percent":19.148936170212767,"passes":27,"runs":138,"runtime":252612.16767463155,"source_flags":[],"source_harness":"ale_claw","source_row_index":687,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":923.69,"totalDurationS":252612.16767463155,"totalInputTokens":1021079941,"totalOutputTokens":5088971,"total_cost_usd":923.69,"total_input_tokens":1021079941,"total_output_tokens":5088971,"total_runtime_s":252612.16767463155,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":41.973334122411735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-88d57d215b1609b15511","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0db8e51118ed5752ea830daa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3521506822590183,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":556,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[277];split=full/overall;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":556,"metadata":{"avgScore":0.3521506822590183,"avg_score":0.3521506822590183,"avg_score_percent":35.21506822590183,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1688.85,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":807329919,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":3963995.5,"passRate":0.15789473684210525,"pass_rate":0.15789473684210525,"pass_rate_percent":15.789473684210526,"passes":24,"runs":142,"runtime":446984.7452694551,"source_flags":[],"source_harness":"openclaw","source_row_index":277,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1688.85,"totalDurationS":446984.7452694551,"totalInputTokens":807329919,"totalOutputTokens":3963995.5,"total_cost_usd":1688.85,"total_input_tokens":807329919,"total_output_tokens":3963995.5,"total_runtime_s":446984.7452694551,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.21506822590183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-602fbc3eca2bf5d30618","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-07cbddf456c798095db5ab9b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4049605336355299,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":64,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[31];split=full;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":64,"metadata":{"avgScore":0.4049605336355299,"avg_score":0.4049605336355299,"avg_score_percent":40.49605336355299,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1132.24,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1342066095,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":5568247,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":143,"runtime":279527.28043104644,"source_flags":[],"source_harness":"ale_claw","source_row_index":31,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1132.24,"totalDurationS":279527.28043104644,"totalInputTokens":1342066095,"totalOutputTokens":5568247,"total_cost_usd":1132.24,"total_input_tokens":1342066095,"total_output_tokens":5568247,"total_runtime_s":279527.28043104644,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":40.49605336355299,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e2110753675e121b5f03","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-01ed28738507d248a47372a4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4197333412241174,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9889192dc3f4a7a00a0af089788c315d3a07059","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":904,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[451];split=unlicensed;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":904,"metadata":{"avgScore":0.4197333412241174,"avg_score":0.4197333412241174,"avg_score_percent":41.973334122411735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":923.69,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1021079941,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-7","output_tokens":5088971,"passRate":0.19148936170212766,"pass_rate":0.19148936170212766,"pass_rate_percent":19.148936170212767,"passes":27,"runs":138,"runtime":252612.16767463155,"source_flags":[],"source_harness":"ale_claw","source_row_index":451,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":923.69,"totalDurationS":252612.16767463155,"totalInputTokens":1021079941,"totalOutputTokens":5088971,"total_cost_usd":923.69,"total_input_tokens":1021079941,"total_output_tokens":5088971,"total_runtime_s":252612.16767463155,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":41.973334122411735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-748c91e14e7fd055c36e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f6eb2640d9fe40b70dc85aea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1153,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[576];split=unlicensed/last-exam;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1153,"metadata":{"avgScore":0.08321514450874437,"avg_score":0.08321514450874437,"avg_score_percent":8.321514450874437,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":425.56,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":450451674,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1846704,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":28,"runtime":135702.5551899186,"source_flags":[],"source_harness":"ale_claw","source_row_index":576,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":425.56,"totalDurationS":135702.5551899186,"totalInputTokens":450451674,"totalOutputTokens":1846704,"total_cost_usd":425.56,"total_input_tokens":450451674,"total_output_tokens":1846704,"total_runtime_s":135702.5551899186,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-28d3372b63e01e1b6b5c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f651c91fd62dc65e8309c7f6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18421052631578946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":63,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[31];split=full;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":63,"metadata":{"avgScore":0.4049605336355299,"avg_score":0.4049605336355299,"avg_score_percent":40.49605336355299,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1132.24,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1342066095,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":5568247,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":143,"runtime":279527.28043104644,"source_flags":[],"source_harness":"ale_claw","source_row_index":31,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1132.24,"totalDurationS":279527.28043104644,"totalInputTokens":1342066095,"totalOutputTokens":5568247,"total_cost_usd":1132.24,"total_input_tokens":1342066095,"total_output_tokens":5568247,"total_runtime_s":279527.28043104644,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.421052631578945,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cf2f54668ddce795f8e0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f640bfd24a49ea9dd1ab4722","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.15789473684210525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":83,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[41];split=full;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":83,"metadata":{"avgScore":0.3521506822590183,"avg_score":0.3521506822590183,"avg_score_percent":35.21506822590183,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1688.85,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":807329919,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3963995.5,"passRate":0.15789473684210525,"pass_rate":0.15789473684210525,"pass_rate_percent":15.789473684210526,"passes":24,"runs":142,"runtime":446984.7452694551,"source_flags":[],"source_harness":"openclaw","source_row_index":41,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1688.85,"totalDurationS":446984.7452694551,"totalInputTokens":807329919,"totalOutputTokens":3963995.5,"total_cost_usd":1688.85,"total_input_tokens":807329919,"total_output_tokens":3963995.5,"total_runtime_s":446984.7452694551,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.789473684210526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-77a061966319a583ae9d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-eb5a7a5c74a9e64c7403359a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.14893617021276595,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":919,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[459];split=unlicensed;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":919,"metadata":{"avgScore":0.3722051359473687,"avg_score":0.3722051359473687,"avg_score_percent":37.22051359473687,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1558.83,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":380178747,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3444699,"passRate":0.14893617021276595,"pass_rate":0.14893617021276595,"pass_rate_percent":14.893617021276595,"passes":21,"runs":130,"runtime":141696.4445367267,"source_flags":[],"source_harness":"claude_code","source_row_index":459,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1558.83,"totalDurationS":141696.4445367267,"totalInputTokens":380178747,"totalOutputTokens":3444699,"total_cost_usd":1558.83,"total_input_tokens":380178747,"total_output_tokens":3444699,"total_runtime_s":141696.4445367267,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.893617021276595,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fef555363cc069b85235","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e83a67a3b7db909b0fdd9e94","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1151,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[575];split=unlicensed/last-exam;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1151,"metadata":{"avgScore":0.08531923866916752,"avg_score":0.08531923866916752,"avg_score_percent":8.531923866916753,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":475.55,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":141000329,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":865873,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":17,"runtime":35455.30997874378,"source_flags":[],"source_harness":"droid","source_row_index":575,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":17,"tokens":null,"totalCostUsd":475.55,"totalDurationS":35455.30997874378,"totalInputTokens":141000329,"totalOutputTokens":865873,"total_cost_usd":475.55,"total_input_tokens":141000329,"total_output_tokens":865873,"total_runtime_s":35455.30997874378,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4e8e6373064bf85ea313","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e51541cf8797c662905e2ec3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22695035460992907,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1351,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[675];split=unlicensed/overall;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1351,"metadata":{"avgScore":0.44713736097477313,"avg_score":0.44713736097477313,"avg_score_percent":44.713736097477316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1564.2,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":317070704,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3915017,"passRate":0.22695035460992907,"pass_rate":0.22695035460992907,"pass_rate_percent":22.69503546099291,"passes":32,"runs":130,"runtime":209777.7502665911,"source_flags":[],"source_harness":"cursor_cli","source_row_index":675,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1564.2,"totalDurationS":209777.7502665911,"totalInputTokens":317070704,"totalOutputTokens":3915017,"total_cost_usd":1564.2,"total_input_tokens":317070704,"total_output_tokens":3915017,"total_runtime_s":209777.7502665911,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.69503546099291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-de1c549f75d68aac5a4a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e49605fa9a8a01dea438bfc9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1453900709219858,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1397,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[698];split=unlicensed/overall;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1397,"metadata":{"avgScore":0.33663835749551463,"avg_score":0.33663835749551463,"avg_score_percent":33.66383574955146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1325.06,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":344833331.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2642831,"passRate":0.1453900709219858,"pass_rate":0.1453900709219858,"pass_rate_percent":14.539007092198581,"passes":22,"runs":101,"runtime":99274.81146285753,"source_flags":[],"source_harness":"droid","source_row_index":698,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":95,"tokens":null,"totalCostUsd":1325.06,"totalDurationS":99274.81146285753,"totalInputTokens":344833331.5,"totalOutputTokens":2642831,"total_cost_usd":1325.06,"total_input_tokens":344833331.5,"total_output_tokens":2642831,"total_runtime_s":99274.81146285753,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.539007092198581,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-622cb8c65ba74ac2a06e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-de72143a8183168fb1cdbbe2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":767,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[383];split=linux_only;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":767,"metadata":{"avgScore":0.4326975564273962,"avg_score":0.4326975564273962,"avg_score_percent":43.26975564273962,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.55,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":389612585,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2906411,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":21,"runs":103,"runtime":159298.15374917473,"source_flags":[],"source_harness":"ale_claw","source_row_index":383,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":415.55,"totalDurationS":159298.15374917473,"totalInputTokens":389612585,"totalOutputTokens":2906411,"total_cost_usd":415.55,"total_input_tokens":389612585,"total_output_tokens":2906411,"total_runtime_s":159298.15374917473,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fc515990312b1c4bfb3d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d403adb4f58b52aa86fc0aa8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.14893617021276595,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1391,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[695];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1391,"metadata":{"avgScore":0.3722051359473687,"avg_score":0.3722051359473687,"avg_score_percent":37.22051359473687,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1558.83,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":380178747,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3444699,"passRate":0.14893617021276595,"pass_rate":0.14893617021276595,"pass_rate_percent":14.893617021276595,"passes":21,"runs":130,"runtime":141696.4445367267,"source_flags":[],"source_harness":"claude_code","source_row_index":695,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1558.83,"totalDurationS":141696.4445367267,"totalInputTokens":380178747,"totalOutputTokens":3444699,"total_cost_usd":1558.83,"total_input_tokens":380178747,"total_output_tokens":3444699,"total_runtime_s":141696.4445367267,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.893617021276595,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b5c39ec53349775733f5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d10480624a6b1c957f7987a3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1011,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[505];split=unlicensed/full-spectrum;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1011,"metadata":{"avgScore":0.36513474394967993,"avg_score":0.36513474394967993,"avg_score_percent":36.51347439496799,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":201.89,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":222692155,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1493759,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":49,"runtime":47608.04235754142,"source_flags":[],"source_harness":"ale_claw","source_row_index":505,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":201.89,"totalDurationS":47608.04235754142,"totalInputTokens":222692155,"totalOutputTokens":1493759,"total_cost_usd":201.89,"total_input_tokens":222692155,"total_output_tokens":1493759,"total_runtime_s":47608.04235754142,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dbfc1c1ce3983feadffb","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d0448e8c2ea894e0b632e866","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.15789473684210525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":555,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[277];split=full/overall;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":555,"metadata":{"avgScore":0.3521506822590183,"avg_score":0.3521506822590183,"avg_score_percent":35.21506822590183,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1688.85,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":807329919,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3963995.5,"passRate":0.15789473684210525,"pass_rate":0.15789473684210525,"pass_rate_percent":15.789473684210526,"passes":24,"runs":142,"runtime":446984.7452694551,"source_flags":[],"source_harness":"openclaw","source_row_index":277,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1688.85,"totalDurationS":446984.7452694551,"totalInputTokens":807329919,"totalOutputTokens":3963995.5,"total_cost_usd":1688.85,"total_input_tokens":807329919,"total_output_tokens":3963995.5,"total_runtime_s":446984.7452694551,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.789473684210526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3cebbf9f371fa2e41932","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cc0e3c74147c769a97f8e49c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":995,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[497];split=unlicensed/full-spectrum;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":995,"metadata":{"avgScore":0.39292900253171503,"avg_score":0.39292900253171503,"avg_score_percent":39.2929002531715,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":580.62,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":128052474,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1427847,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":48,"runtime":40744.41926420876,"source_flags":[],"source_harness":"cursor_cli","source_row_index":497,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":580.62,"totalDurationS":40744.41926420876,"totalInputTokens":128052474,"totalOutputTokens":1427847,"total_cost_usd":580.62,"total_input_tokens":128052474,"total_output_tokens":1427847,"total_runtime_s":40744.41926420876,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-736b97ad9d4e9989d82f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c7d64699e1ac29ead11555fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":651,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[325];split=licensed;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":651,"metadata":{"avgScore":0.13946830391958281,"avg_score":0.13946830391958281,"avg_score_percent":13.946830391958281,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":335.13,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":112444567,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":549497,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":19954.802037291927,"source_flags":[],"source_harness":"cursor_cli","source_row_index":325,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":335.13,"totalDurationS":19954.802037291927,"totalInputTokens":112444567,"totalOutputTokens":549497,"total_cost_usd":335.13,"total_input_tokens":112444567,"total_output_tokens":549497,"total_runtime_s":19954.802037291927,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-906770a0704bcc909156","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c70c396926c045f60397a483","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.12727272727272726,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":197,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[98];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":197,"metadata":{"avgScore":0.29108336390143064,"avg_score":0.29108336390143064,"avg_score_percent":29.108336390143062,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":747.4,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":201998860,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1546037,"passRate":0.12727272727272726,"pass_rate":0.12727272727272726,"pass_rate_percent":12.727272727272727,"passes":7,"runs":50,"runtime":43471.418552602874,"source_flags":[],"source_harness":"claude_code","source_row_index":98,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":747.4,"totalDurationS":43471.418552602874,"totalInputTokens":201998860,"totalOutputTokens":1546037,"total_cost_usd":747.4,"total_input_tokens":201998860,"total_output_tokens":1546037,"total_runtime_s":43471.418552602874,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.727272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f7811f110bcc73ef885f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c702ffa1e2ed9324310c5148","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22857142857142856,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":751,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[375];split=linux_only;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":751,"metadata":{"avgScore":0.45680671377661547,"avg_score":0.45680671377661547,"avg_score_percent":45.68067137766155,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1044.57,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":207544892,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2506674,"passRate":0.22857142857142856,"pass_rate":0.22857142857142856,"pass_rate_percent":22.857142857142858,"passes":24,"runs":96,"runtime":89632.5721885981,"source_flags":[],"source_harness":"cursor_cli","source_row_index":375,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":96,"tokens":null,"totalCostUsd":1044.57,"totalDurationS":89632.5721885981,"totalInputTokens":207544892,"totalOutputTokens":2506674,"total_cost_usd":1044.57,"total_input_tokens":207544892,"total_output_tokens":2506674,"total_runtime_s":89632.5721885981,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.857142857142858,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-31565c09550fc18e4440","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bf7a417cf0895a3b1d5c62c2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.31343283582089554,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":405,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[202];split=full/near-term;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":405,"metadata":{"avgScore":0.6269364006481253,"avg_score":0.6269364006481253,"avg_score_percent":62.69364006481253,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":557.68,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":110261629,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1648119,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":120381.51676356932,"source_flags":[],"source_harness":"cursor_cli","source_row_index":202,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":557.68,"totalDurationS":120381.51676356932,"totalInputTokens":110261629,"totalOutputTokens":1648119,"total_cost_usd":557.68,"total_input_tokens":110261629,"total_output_tokens":1648119,"total_runtime_s":120381.51676356932,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.343283582089555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eb3b1ff0bdb193246538","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bccf3d77219afd625a5e98c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.31343283582089554,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1241,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[620];split=unlicensed/near-term;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1241,"metadata":{"avgScore":0.6269364006481253,"avg_score":0.6269364006481253,"avg_score_percent":62.69364006481253,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":557.68,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":110261629,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1648119,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":120381.51676356932,"source_flags":[],"source_harness":"cursor_cli","source_row_index":620,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":557.68,"totalDurationS":120381.51676356932,"totalInputTokens":110261629,"totalOutputTokens":1648119,"total_cost_usd":557.68,"total_input_tokens":110261629,"total_output_tokens":1648119,"total_runtime_s":120381.51676356932,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.343283582089555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e06c337071b630633e21","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b8e1f8410b495e66f8508f40","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21052631578947367,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":519,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[259];split=full/overall;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":519,"metadata":{"avgScore":0.42487183710893694,"avg_score":0.42487183710893694,"avg_score_percent":42.487183710893696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1899.32,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":429515271,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":4464514,"passRate":0.21052631578947367,"pass_rate":0.21052631578947367,"pass_rate_percent":21.052631578947366,"passes":32,"runs":137,"runtime":229732.552303883,"source_flags":[],"source_harness":"cursor_cli","source_row_index":259,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1899.32,"totalDurationS":229732.552303883,"totalInputTokens":429515271,"totalOutputTokens":4464514,"total_cost_usd":1899.32,"total_input_tokens":429515271,"total_output_tokens":4464514,"total_runtime_s":229732.552303883,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.052631578947366,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2c7c4f6ce9ed0e5e9018","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b5406f16fb5d4a00810d8ed0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19148936170212766,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1375,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[687];split=unlicensed/overall;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1375,"metadata":{"avgScore":0.4197333412241174,"avg_score":0.4197333412241174,"avg_score_percent":41.973334122411735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":923.69,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1021079941,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":5088971,"passRate":0.19148936170212766,"pass_rate":0.19148936170212766,"pass_rate_percent":19.148936170212767,"passes":27,"runs":138,"runtime":252612.16767463155,"source_flags":[],"source_harness":"ale_claw","source_row_index":687,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":923.69,"totalDurationS":252612.16767463155,"totalInputTokens":1021079941,"totalOutputTokens":5088971,"total_cost_usd":923.69,"total_input_tokens":1021079941,"total_output_tokens":5088971,"total_runtime_s":252612.16767463155,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.148936170212767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8c42b2024c6268ef5db6","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ae99273287f738967794227a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":433,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[216];split=full/near-term;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":433,"metadata":{"avgScore":0.5799257416141242,"avg_score":0.5799257416141242,"avg_score_percent":57.99257416141243,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":508.37,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":201660470,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1516121.5,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":66,"runtime":171722.06837378268,"source_flags":[],"source_harness":"openclaw","source_row_index":216,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":508.37,"totalDurationS":171722.06837378268,"totalInputTokens":201660470,"totalOutputTokens":1516121.5,"total_cost_usd":508.37,"total_input_tokens":201660470,"total_output_tokens":1516121.5,"total_runtime_s":171722.06837378268,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ffa55dbbb1b780022cfd","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ac7828957a7a04e4cf66b261","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":167,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[83];split=full/full-spectrum;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":167,"metadata":{"avgScore":0.37616366308547566,"avg_score":0.37616366308547566,"avg_score_percent":37.61636630854757,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":841.57,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":215576200,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1889131,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":53,"runtime":53276.67023467668,"source_flags":[],"source_harness":"cursor_cli","source_row_index":83,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":841.57,"totalDurationS":53276.67023467668,"totalInputTokens":215576200,"totalOutputTokens":1889131,"total_cost_usd":841.57,"total_input_tokens":215576200,"total_output_tokens":1889131,"total_runtime_s":53276.67023467668,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0e2e3d2f3b92073b25f3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a5eece0389dcfc883a01dba8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1269,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[634];split=unlicensed/near-term;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1269,"metadata":{"avgScore":0.5799257416141242,"avg_score":0.5799257416141242,"avg_score_percent":57.99257416141243,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":508.37,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":201660470,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1516121.5,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":66,"runtime":171722.06837378268,"source_flags":[],"source_harness":"openclaw","source_row_index":634,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":508.37,"totalDurationS":171722.06837378268,"totalInputTokens":201660470,"totalOutputTokens":1516121.5,"total_cost_usd":508.37,"total_input_tokens":201660470,"total_output_tokens":1516121.5,"total_runtime_s":171722.06837378268,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c056bf64a51555aa1d97","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a3c8f34bced9cedb5a21ed1f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.1523809523809524,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":789,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[394];split=linux_only;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":789,"metadata":{"avgScore":0.3890574125005675,"avg_score":0.3890574125005675,"avg_score_percent":38.90574125005675,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1164.8,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":280978346,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2723233,"passRate":0.1523809523809524,"pass_rate":0.1523809523809524,"pass_rate_percent":15.238095238095239,"passes":16,"runs":96,"runtime":92839.4148778622,"source_flags":[],"source_harness":"claude_code","source_row_index":394,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":96,"tokens":null,"totalCostUsd":1164.8,"totalDurationS":92839.4148778622,"totalInputTokens":280978346,"totalOutputTokens":2723233,"total_cost_usd":1164.8,"total_input_tokens":280978346,"total_output_tokens":2723233,"total_runtime_s":92839.4148778622,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":15.238095238095239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3f9b8e6e58b260585c5c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a2c469b737f95178f259d67d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":681,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[340];split=licensed;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":681,"metadata":{"avgScore":0.06468889570193674,"avg_score":0.06468889570193674,"avg_score_percent":6.468889570193674,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.19,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":7226933,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":83461,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":4,"runtime":2505.957411190262,"source_flags":[],"source_harness":"droid","source_row_index":340,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":4,"tokens":null,"totalCostUsd":31.19,"totalDurationS":2505.957411190262,"totalInputTokens":7226933,"totalOutputTokens":83461,"total_cost_usd":31.19,"total_input_tokens":7226933,"total_output_tokens":83461,"total_runtime_s":2505.957411190262,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d162935e0a42b66523dc","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a13aaccb80dc4b728454acfd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":173,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[86];split=full/full-spectrum;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":173,"metadata":{"avgScore":0.3661224944997091,"avg_score":0.3661224944997091,"avg_score_percent":36.61224944997091,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.46,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":288810311,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1669212,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":52,"runtime":55840.03182199924,"source_flags":[],"source_harness":"ale_claw","source_row_index":86,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":52,"tokens":null,"totalCostUsd":246.46,"totalDurationS":55840.03182199924,"totalInputTokens":288810311,"totalOutputTokens":1669212,"total_cost_usd":246.46,"total_input_tokens":288810311,"total_output_tokens":1669212,"total_runtime_s":55840.03182199924,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e3062dc7d860cbd35a3e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9f9bafb52046ded5a83253ad","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1523809523809524,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":795,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[397];split=linux_only;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":795,"metadata":{"avgScore":0.3469339382470838,"avg_score":0.3469339382470838,"avg_score_percent":34.69339382470838,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":709.97,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":176866012,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1651850,"passRate":0.1523809523809524,"pass_rate":0.1523809523809524,"pass_rate_percent":15.238095238095239,"passes":17,"runs":74,"runtime":69448.19969177898,"source_flags":[],"source_harness":"droid","source_row_index":397,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":70,"tokens":null,"totalCostUsd":709.97,"totalDurationS":69448.19969177898,"totalInputTokens":176866012,"totalOutputTokens":1651850,"total_cost_usd":709.97,"total_input_tokens":176866012,"total_output_tokens":1651850,"total_runtime_s":69448.19969177898,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.238095238095239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a07b3eaf97af850142a1","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9c2d8eed4ace469777ff9c25","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19148936170212766,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":903,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[451];split=unlicensed;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":903,"metadata":{"avgScore":0.4197333412241174,"avg_score":0.4197333412241174,"avg_score_percent":41.973334122411735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":923.69,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1021079941,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":5088971,"passRate":0.19148936170212766,"pass_rate":0.19148936170212766,"pass_rate_percent":19.148936170212767,"passes":27,"runs":138,"runtime":252612.16767463155,"source_flags":[],"source_harness":"ale_claw","source_row_index":451,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":923.69,"totalDurationS":252612.16767463155,"totalInputTokens":1021079941,"totalOutputTokens":5088971,"total_cost_usd":923.69,"total_input_tokens":1021079941,"total_output_tokens":5088971,"total_runtime_s":252612.16767463155,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.148936170212767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b7f761aa7ee9a2045ba2","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-97f21fc645d46e14aeab61f2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1109,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[554];split=unlicensed/last-exam;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1109,"metadata":{"avgScore":0.12847049593392926,"avg_score":0.12847049593392926,"avg_score_percent":12.847049593392926,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":496.57,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":92176971,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1046938,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":22,"runtime":58179.24658737797,"source_flags":[],"source_harness":"cursor_cli","source_row_index":554,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":22,"tokens":null,"totalCostUsd":496.57,"totalDurationS":58179.24658737797,"totalInputTokens":92176971,"totalOutputTokens":1046938,"total_cost_usd":496.57,"total_input_tokens":92176971,"total_output_tokens":1046938,"total_runtime_s":58179.24658737797,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b473ddea6d7832be1bb1","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-91dc63b4c3ecdf4de711e876","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1255,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[627];split=unlicensed/near-term;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1255,"metadata":{"avgScore":0.6045069241062173,"avg_score":0.6045069241062173,"avg_score_percent":60.450692410621734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":311.72,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":359152755,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1944556,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":74416.76165142255,"source_flags":[],"source_harness":"ale_claw","source_row_index":627,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":311.72,"totalDurationS":74416.76165142255,"totalInputTokens":359152755,"totalOutputTokens":1944556,"total_cost_usd":311.72,"total_input_tokens":359152755,"total_output_tokens":1944556,"total_runtime_s":74416.76165142255,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1efd6d1c48faa1b975ff","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8f572318317059429b7a0205","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1037,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[518];split=unlicensed/full-spectrum;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1037,"metadata":{"avgScore":0.2777319707686421,"avg_score":0.2777319707686421,"avg_score_percent":27.773197076864207,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":301.67,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":126991014,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1028842,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":5,"runs":48,"runtime":81810.31944376754,"source_flags":[],"source_harness":"openclaw","source_row_index":518,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":301.67,"totalDurationS":81810.31944376754,"totalInputTokens":126991014,"totalOutputTokens":1028842,"total_cost_usd":301.67,"total_input_tokens":126991014,"total_output_tokens":1028842,"total_runtime_s":81810.31944376754,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ec6ae43a300b48177e13","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8b20358df211cc9592930198","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":301,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[150];split=full/last-exam;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":301,"metadata":{"avgScore":0.10029435149513931,"avg_score":0.10029435149513931,"avg_score_percent":10.029435149513931,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":684.93,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":170082318,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1387243,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":61575.264092724305,"source_flags":[],"source_harness":"claude_code","source_row_index":150,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":684.93,"totalDurationS":61575.264092724305,"totalInputTokens":170082318,"totalOutputTokens":1387243,"total_cost_usd":684.93,"total_input_tokens":170082318,"total_output_tokens":1387243,"total_runtime_s":61575.264092724305,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f6878a5daeead5dfea1b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8692335995b2258ebad0e6e4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":333,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[166];split=full/last-exam;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":333,"metadata":{"avgScore":0.043412991868513016,"avg_score":0.043412991868513016,"avg_score_percent":4.341299186851302,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":841.65,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":452898006,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1462721,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":32,"runtime":180025.200925768,"source_flags":[],"source_harness":"openclaw","source_row_index":166,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":841.65,"totalDurationS":180025.200925768,"totalInputTokens":452898006,"totalOutputTokens":1462721,"total_cost_usd":841.65,"total_input_tokens":452898006,"total_output_tokens":1462721,"total_runtime_s":180025.200925768,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-76b246489b88fb13c070","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7c53603edaf92598fa9eaabf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":315,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[157];split=full/last-exam;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":315,"metadata":{"avgScore":0.08029413579144805,"avg_score":0.08029413579144805,"avg_score_percent":8.029413579144805,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":495.79,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":145716643,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":920145,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":20,"runtime":37053.47873265087,"source_flags":[],"source_harness":"droid","source_row_index":157,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":20,"tokens":null,"totalCostUsd":495.79,"totalDurationS":37053.47873265087,"totalInputTokens":145716643,"totalOutputTokens":920145,"total_cost_usd":495.79,"total_input_tokens":145716643,"total_output_tokens":920145,"total_runtime_s":37053.47873265087,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1d54da48213ed6d6692b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-79638ca316e175d916955eb1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":419,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[209];split=full/near-term;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":419,"metadata":{"avgScore":0.6045069241062173,"avg_score":0.6045069241062173,"avg_score_percent":60.450692410621734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":311.72,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":359152755,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1944556,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":74416.76165142255,"source_flags":[],"source_harness":"ale_claw","source_row_index":209,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":311.72,"totalDurationS":74416.76165142255,"totalInputTokens":359152755,"totalOutputTokens":1944556,"total_cost_usd":311.72,"total_input_tokens":359152755,"total_output_tokens":1944556,"total_runtime_s":74416.76165142255,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-66cbe71cb1a92b2df671","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7330baf1face344817cc8b76","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.02,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1067,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[533];split=unlicensed/full-spectrum;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1067,"metadata":{"avgScore":0.09501958811261163,"avg_score":0.09501958811261163,"avg_score_percent":9.501958811261163,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":150.16,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":36489552,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":423717,"passRate":0.02,"pass_rate":0.02,"pass_rate_percent":2.0,"passes":1,"runs":18,"runtime":10051.606426970102,"source_flags":[],"source_harness":"droid","source_row_index":533,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":18,"tokens":null,"totalCostUsd":150.16,"totalDurationS":10051.606426970102,"totalInputTokens":36489552,"totalOutputTokens":423717,"total_cost_usd":150.16,"total_input_tokens":36489552,"total_output_tokens":423717,"total_runtime_s":10051.606426970102,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a9b33622793890d2e00b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6d3e816c62a65bd0e2e408ca","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1702127659574468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":917,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[458];split=unlicensed;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":917,"metadata":{"avgScore":0.3745618668551661,"avg_score":0.3745618668551661,"avg_score_percent":37.45618668551661,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1495.3,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":722150808,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3808451.5,"passRate":0.1702127659574468,"pass_rate":0.1702127659574468,"pass_rate_percent":17.02127659574468,"passes":24,"runs":137,"runtime":403839.59668815427,"source_flags":[],"source_harness":"openclaw","source_row_index":458,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1495.3,"totalDurationS":403839.59668815427,"totalInputTokens":722150808,"totalOutputTokens":3808451.5,"total_cost_usd":1495.3,"total_input_tokens":722150808,"total_output_tokens":3808451.5,"total_runtime_s":403839.59668815427,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.02127659574468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-121efebf83fa6cf2170a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-613402515ff4fdcbe9480198","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1147,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[573];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1147,"metadata":{"avgScore":0.08768284482818715,"avg_score":0.08768284482818715,"avg_score_percent":8.768284482818716,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":576.53,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":134441437,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1284319,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":26,"runtime":57084.972001090646,"source_flags":[],"source_harness":"claude_code","source_row_index":573,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":576.53,"totalDurationS":57084.972001090646,"totalInputTokens":134441437,"totalOutputTokens":1284319,"total_cost_usd":576.53,"total_input_tokens":134441437,"total_output_tokens":1284319,"total_runtime_s":57084.972001090646,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c2d4349c49649e998a88","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5e362706077dc9c8b06cff50","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13815789473684212,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":81,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[40];split=full;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":81,"metadata":{"avgScore":0.35803700118782017,"avg_score":0.35803700118782017,"avg_score_percent":35.80370011878202,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1792.72,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":456438963,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3700468,"passRate":0.13815789473684212,"pass_rate":0.13815789473684212,"pass_rate_percent":13.815789473684212,"passes":21,"runs":137,"runtime":153370.83306965837,"source_flags":[],"source_harness":"claude_code","source_row_index":40,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1792.72,"totalDurationS":153370.83306965837,"totalInputTokens":456438963,"totalOutputTokens":3700468,"total_cost_usd":1792.72,"total_input_tokens":456438963,"total_output_tokens":3700468,"total_runtime_s":153370.83306965837,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.815789473684212,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-de0d74472a0c2f3a0822","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5ca7c331a875df8ebefa471f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.22388059701492538,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":437,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[218];split=full/near-term;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":437,"metadata":{"avgScore":0.5578444176897955,"avg_score":0.5578444176897955,"avg_score_percent":55.784441768979555,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":496.33,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":120384368,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1068334,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":63,"runtime":58636.88796496298,"source_flags":[],"source_harness":"claude_code","source_row_index":218,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":496.33,"totalDurationS":58636.88796496298,"totalInputTokens":120384368,"totalOutputTokens":1068334,"total_cost_usd":496.33,"total_input_tokens":120384368,"total_output_tokens":1068334,"total_runtime_s":58636.88796496298,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.388059701492537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d94f42ed28c74e2c9cc8","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-595d6f524ffe1f0e38c51c09","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.22388059701492538,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1273,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[636];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1273,"metadata":{"avgScore":0.5578444176897955,"avg_score":0.5578444176897955,"avg_score_percent":55.784441768979555,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":496.33,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":120384368,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1068334,"passRate":0.22388059701492538,"pass_rate":0.22388059701492538,"pass_rate_percent":22.388059701492537,"passes":15,"runs":63,"runtime":58636.88796496298,"source_flags":[],"source_harness":"claude_code","source_row_index":636,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":496.33,"totalDurationS":58636.88796496298,"totalInputTokens":120384368,"totalOutputTokens":1068334,"total_cost_usd":496.33,"total_input_tokens":120384368,"total_output_tokens":1068334,"total_runtime_s":58636.88796496298,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.388059701492537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-596b0e91a1a4ba741930","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4e53835fafe728b45c85db53","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.17142857142857143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":791,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[395];split=linux_only;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":791,"metadata":{"avgScore":0.38739364141550475,"avg_score":0.38739364141550475,"avg_score_percent":38.739364141550475,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":959.57,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":516245034,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2766369.5,"passRate":0.17142857142857143,"pass_rate":0.17142857142857143,"pass_rate_percent":17.142857142857142,"passes":18,"runs":104,"runtime":192542.93685384613,"source_flags":[],"source_harness":"openclaw","source_row_index":395,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":959.57,"totalDurationS":192542.93685384613,"totalInputTokens":516245034,"totalOutputTokens":2766369.5,"total_cost_usd":959.57,"total_input_tokens":516245034,"total_output_tokens":2766369.5,"total_runtime_s":192542.93685384613,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.142857142857142,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8fce360ac9df9ba10b33","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3939a3d7774e775f676586e3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":203,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[101];split=full/full-spectrum;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":203,"metadata":{"avgScore":0.274703254822263,"avg_score":0.274703254822263,"avg_score_percent":27.4703254822263,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":369.56,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":162969004,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1135215,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":51,"runtime":107204.82072244062,"source_flags":[],"source_harness":"openclaw","source_row_index":101,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":51,"tokens":null,"totalCostUsd":369.56,"totalDurationS":107204.82072244062,"totalInputTokens":162969004,"totalOutputTokens":1135215,"total_cost_usd":369.56,"total_input_tokens":162969004,"total_output_tokens":1135215,"total_runtime_s":107204.82072244062,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.909090909090908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eb27b6e362321f474df0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-345b79e247067776fae30b66","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":623,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[311];split=licensed;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":623,"metadata":{"avgScore":0.21559999999999999,"avg_score":0.21559999999999999,"avg_score_percent":21.56,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":208.55,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":320986154,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":479276,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":5,"runtime":26915.112756414877,"source_flags":[],"source_harness":"ale_claw","source_row_index":311,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":208.55,"totalDurationS":26915.112756414877,"totalInputTokens":320986154,"totalOutputTokens":479276,"total_cost_usd":208.55,"total_input_tokens":320986154,"total_output_tokens":479276,"total_runtime_s":26915.112756414877,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eebd5e280bb49568add4","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-33ca4f9280179f7e3acebb0a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1702127659574468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1389,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[694];split=unlicensed/overall;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1389,"metadata":{"avgScore":0.3745618668551661,"avg_score":0.3745618668551661,"avg_score_percent":37.45618668551661,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1495.3,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":722150808,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3808451.5,"passRate":0.1702127659574468,"pass_rate":0.1702127659574468,"pass_rate_percent":17.02127659574468,"passes":24,"runs":137,"runtime":403839.59668815427,"source_flags":[],"source_harness":"openclaw","source_row_index":694,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1495.3,"totalDurationS":403839.59668815427,"totalInputTokens":722150808,"totalOutputTokens":3808451.5,"total_cost_usd":1495.3,"total_input_tokens":722150808,"total_output_tokens":3808451.5,"total_runtime_s":403839.59668815427,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.02127659574468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c9b4fb8505edff241ddf","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-31df4e2e809dad8cc87d56e2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.291044776119403,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1249,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[624];split=unlicensed/near-term;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1249,"metadata":{"avgScore":0.6171275522572685,"avg_score":0.6171275522572685,"avg_score_percent":61.712755225726845,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":737.55,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":176244296.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1470123,"passRate":0.291044776119403,"pass_rate":0.291044776119403,"pass_rate_percent":29.1044776119403,"passes":21,"runs":69,"runtime":57944.460609696805,"source_flags":[],"source_harness":"droid","source_row_index":624,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":737.55,"totalDurationS":57944.460609696805,"totalInputTokens":176244296.5,"totalOutputTokens":1470123,"total_cost_usd":737.55,"total_input_tokens":176244296.5,"total_output_tokens":1470123,"total_runtime_s":57944.460609696805,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.1044776119403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-86da3c2ba7c146c2dc6d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3107ea084e44a17d6a0fca88","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.12,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1039,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[519];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1039,"metadata":{"avgScore":0.27546770029157375,"avg_score":0.27546770029157375,"avg_score_percent":27.546770029157376,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":559.29,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":138650190,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1350638,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":46,"runtime":34153.082216965966,"source_flags":[],"source_harness":"claude_code","source_row_index":519,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":46,"tokens":null,"totalCostUsd":559.29,"totalDurationS":34153.082216965966,"totalInputTokens":138650190,"totalOutputTokens":1350638,"total_cost_usd":559.29,"total_input_tokens":138650190,"total_output_tokens":1350638,"total_runtime_s":34153.082216965966,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cfdba5c2a00c24bf8360","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2bfc46cd50a785e393c870fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13486842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":89,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[44];split=full;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":89,"metadata":{"avgScore":0.31695780433940046,"avg_score":0.31695780433940046,"avg_score_percent":31.695780433940048,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1356.25,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":352060264.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2726292,"passRate":0.13486842105263158,"pass_rate":0.13486842105263158,"pass_rate_percent":13.486842105263158,"passes":22,"runs":105,"runtime":101780.76887404779,"source_flags":[],"source_harness":"droid","source_row_index":44,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":99,"tokens":null,"totalCostUsd":1356.25,"totalDurationS":101780.76887404779,"totalInputTokens":352060264.5,"totalOutputTokens":2726292,"total_cost_usd":1356.25,"total_input_tokens":352060264.5,"total_output_tokens":2726292,"total_runtime_s":101780.76887404779,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.486842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-54fbcb8e398a0405d904","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2b05d367e690bc666651a99a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.291044776119403,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":413,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[206];split=full/near-term;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":413,"metadata":{"avgScore":0.6171275522572685,"avg_score":0.6171275522572685,"avg_score_percent":61.712755225726845,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":737.55,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":176244296.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1470123,"passRate":0.291044776119403,"pass_rate":0.291044776119403,"pass_rate_percent":29.1044776119403,"passes":21,"runs":69,"runtime":57944.460609696805,"source_flags":[],"source_harness":"droid","source_row_index":206,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":737.55,"totalDurationS":57944.460609696805,"totalInputTokens":176244296.5,"totalOutputTokens":1470123,"total_cost_usd":737.55,"total_input_tokens":176244296.5,"total_output_tokens":1470123,"total_runtime_s":57944.460609696805,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.1044776119403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-73e855156ec9fbf5692e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2a1b94c6c138963d62fae0a6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":641,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[320];split=licensed;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":641,"metadata":{"avgScore":0.17642727381542536,"avg_score":0.17642727381542536,"avg_score_percent":17.642727381542535,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":233.89,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":76260216,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":255769,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":11674.388532931684,"source_flags":[],"source_harness":"claude_code","source_row_index":320,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":233.89,"totalDurationS":11674.388532931684,"totalInputTokens":76260216,"totalOutputTokens":255769,"total_cost_usd":233.89,"total_input_tokens":76260216,"total_output_tokens":255769,"total_runtime_s":11674.388532931684,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-490cc80beaf9c2eab351","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-26b3af1db026d3c33bd918cc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":281,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[140];split=full/last-exam;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":281,"metadata":{"avgScore":0.11436091784257574,"avg_score":0.11436091784257574,"avg_score_percent":11.436091784257574,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":587.73,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":121011885,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1184118,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":26,"runtime":66683.20773970685,"source_flags":[],"source_harness":"cursor_cli","source_row_index":140,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":26,"tokens":null,"totalCostUsd":587.73,"totalDurationS":66683.20773970685,"totalInputTokens":121011885,"totalOutputTokens":1184118,"total_cost_usd":587.73,"total_input_tokens":121011885,"total_output_tokens":1184118,"total_runtime_s":66683.20773970685,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0e47db07a35865b0f94f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2628925667b577f1586924bc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13486842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":561,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[280];split=full/overall;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":561,"metadata":{"avgScore":0.31695780433940046,"avg_score":0.31695780433940046,"avg_score_percent":31.695780433940048,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1356.25,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":352060264.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2726292,"passRate":0.13486842105263158,"pass_rate":0.13486842105263158,"pass_rate_percent":13.486842105263158,"passes":22,"runs":105,"runtime":101780.76887404779,"source_flags":[],"source_harness":"droid","source_row_index":280,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":99,"tokens":null,"totalCostUsd":1356.25,"totalDurationS":101780.76887404779,"totalInputTokens":352060264.5,"totalOutputTokens":2726292,"total_cost_usd":1356.25,"total_input_tokens":352060264.5,"total_output_tokens":2726292,"total_runtime_s":101780.76887404779,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.486842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-60691ae2cc30653fcfd4","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-222ae5bec5992bf5b3e9e073","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22695035460992907,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":879,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[439];split=unlicensed;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":879,"metadata":{"avgScore":0.44713736097477313,"avg_score":0.44713736097477313,"avg_score_percent":44.713736097477316,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1564.2,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":317070704,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3915017,"passRate":0.22695035460992907,"pass_rate":0.22695035460992907,"pass_rate_percent":22.69503546099291,"passes":32,"runs":130,"runtime":209777.7502665911,"source_flags":[],"source_harness":"cursor_cli","source_row_index":439,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":130,"tokens":null,"totalCostUsd":1564.2,"totalDurationS":209777.7502665911,"totalInputTokens":317070704,"totalOutputTokens":3915017,"total_cost_usd":1564.2,"total_input_tokens":317070704,"total_output_tokens":3915017,"total_runtime_s":209777.7502665911,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.69503546099291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b733a305e5ab216dd409","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1d5dfd94ee563356a76531e3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"thinking-high","source_harness":"cursor_cli","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21052631578947367,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":47,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[23];split=full;harness=cursor_cli;model=claude-opus-4-7;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":47,"metadata":{"avgScore":0.42487183710893694,"avg_score":0.42487183710893694,"avg_score_percent":42.487183710893696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1899.32,"costSource":"official+adj","cost_source":"official+adj","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":"thinking-high","input_tokens":429515271,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":4464514,"passRate":0.21052631578947367,"pass_rate":0.21052631578947367,"pass_rate_percent":21.052631578947366,"passes":32,"runs":137,"runtime":229732.552303883,"source_flags":[],"source_harness":"cursor_cli","source_row_index":23,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1899.32,"totalDurationS":229732.552303883,"totalInputTokens":429515271,"totalOutputTokens":4464514,"total_cost_usd":1899.32,"total_input_tokens":429515271,"total_output_tokens":4464514,"total_runtime_s":229732.552303883,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.052631578947366,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0c80204486bca5d5e388","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1af492d09f16d9b05fedad7d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1453900709219858,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":925,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[462];split=unlicensed;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":925,"metadata":{"avgScore":0.33663835749551463,"avg_score":0.33663835749551463,"avg_score_percent":33.66383574955146,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1325.06,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":344833331.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2642831,"passRate":0.1453900709219858,"pass_rate":0.1453900709219858,"pass_rate_percent":14.539007092198581,"passes":22,"runs":101,"runtime":99274.81146285753,"source_flags":[],"source_harness":"droid","source_row_index":462,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":95,"tokens":null,"totalCostUsd":1325.06,"totalDurationS":99274.81146285753,"totalInputTokens":344833331.5,"totalOutputTokens":2642831,"total_cost_usd":1325.06,"total_input_tokens":344833331.5,"total_output_tokens":2642831,"total_runtime_s":99274.81146285753,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.539007092198581,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-48eb419833869d761cba","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-15a59a95f176ed3f8b70a96e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13815789473684212,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":553,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[276];split=full/overall;harness=claude_code;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":553,"metadata":{"avgScore":0.35803700118782017,"avg_score":0.35803700118782017,"avg_score_percent":35.80370011878202,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1792.72,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":456438963,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":3700468,"passRate":0.13815789473684212,"pass_rate":0.13815789473684212,"pass_rate_percent":13.815789473684212,"passes":21,"runs":137,"runtime":153370.83306965837,"source_flags":[],"source_harness":"claude_code","source_row_index":276,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":1792.72,"totalDurationS":153370.83306965837,"totalInputTokens":456438963,"totalOutputTokens":3700468,"total_cost_usd":1792.72,"total_input_tokens":456438963,"total_output_tokens":3700468,"total_runtime_s":153370.83306965837,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.815789473684212,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-da553da3b5ac42be713a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-14fa2e0270351adc95313b25","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":317,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[158];split=full/last-exam;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":317,"metadata":{"avgScore":0.07863300882269293,"avg_score":0.07863300882269293,"avg_score_percent":7.863300882269293,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":593.61,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":707313906,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":2192916,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":155269.84613300068,"source_flags":[],"source_harness":"ale_claw","source_row_index":158,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":593.61,"totalDurationS":155269.84613300068,"totalInputTokens":707313906,"totalOutputTokens":2192916,"total_cost_usd":593.61,"total_input_tokens":707313906,"total_output_tokens":2192916,"total_runtime_s":155269.84613300068,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e85a3eadfeacbe17a389","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-14dfa4f99ed3da97a22e3d19","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18421052631578946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":535,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[267];split=full/overall;harness=ale_claw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":535,"metadata":{"avgScore":0.4049605336355299,"avg_score":0.4049605336355299,"avg_score_percent":40.49605336355299,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1132.24,"costSource":"official","cost_source":"official","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1342066095,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":5568247,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":143,"runtime":279527.28043104644,"source_flags":[],"source_harness":"ale_claw","source_row_index":267,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1132.24,"totalDurationS":279527.28043104644,"totalInputTokens":1342066095,"totalOutputTokens":5568247,"total_cost_usd":1132.24,"total_input_tokens":1342066095,"total_output_tokens":5568247,"total_runtime_s":279527.28043104644,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.421052631578945,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-840e0f5a5b49dcfb6dbe","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0f2f5d68c51cba3f827b74dd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":231,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[115];split=full/full-spectrum;harness=droid;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":231,"metadata":{"avgScore":0.10856285924276156,"avg_score":0.10856285924276156,"avg_score_percent":10.856285924276156,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":166.81,"costSource":"official+adj","cost_source":"official+adj","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":39918384,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":472069,"passRate":0.03636363636363636,"pass_rate":0.03636363636363636,"pass_rate_percent":3.6363636363636362,"passes":2,"runs":20,"runtime":11256.058063949225,"source_flags":[],"source_harness":"droid","source_row_index":115,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":20,"tokens":null,"totalCostUsd":166.81,"totalDurationS":11256.058063949225,"totalInputTokens":39918384,"totalOutputTokens":472069,"total_cost_usd":166.81,"total_input_tokens":39918384,"total_output_tokens":472069,"total_runtime_s":11256.058063949225,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.6363636363636362,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5d5641ec8e9e8fe63d28","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-01adb0bd392af146215fb885","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1173,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[586];split=unlicensed/last-exam;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1173,"metadata":{"avgScore":0.03860312303344982,"avg_score":0.03860312303344982,"avg_score_percent":3.8603123033449824,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":714.1,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":403184810,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":1398987,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":29,"runtime":160106.59431680728,"source_flags":[],"source_harness":"openclaw","source_row_index":586,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":714.1,"totalDurationS":160106.59431680728,"totalInputTokens":403184810,"totalOutputTokens":1398987,"total_cost_usd":714.1,"total_input_tokens":403184810,"total_output_tokens":1398987,"total_runtime_s":160106.59431680728,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":60,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-72625e9bad32747782e8","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-00f48439416166692256052c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"589d5fad61c0e7c76705a23e9698837d63589dfa","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":677,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[338];split=licensed;harness=openclaw;model=claude-opus-4-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":677,"metadata":{"avgScore":0.06488004334475987,"avg_score":0.06488004334475987,"avg_score_percent":6.4880043344759875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":193.55,"costSource":"official+adj","cost_source":"official+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":85179111,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-7","output_tokens":155544,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":43145.14858130082,"source_flags":[],"source_harness":"openclaw","source_row_index":338,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":193.55,"totalDurationS":43145.14858130082,"totalInputTokens":85179111,"totalOutputTokens":155544,"total_cost_usd":193.55,"total_input_tokens":85179111,"total_output_tokens":155544,"total_runtime_s":43145.14858130082,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3fa4f0b284992ff6b1c1","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-fd02b956f64b4f4b00d03b7d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44189296504626746,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1360,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[679];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1360,"metadata":{"avgScore":0.44189296504626746,"avg_score":0.44189296504626746,"avg_score_percent":44.189296504626746,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2611.28,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1134598082,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":15341352,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":140,"runtime":416112,"source_flags":[],"source_harness":"claude_code","source_row_index":679,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":2611.28,"totalDurationS":416112,"totalInputTokens":1134598082,"totalOutputTokens":15341352,"total_cost_usd":2611.28,"total_input_tokens":1134598082,"total_output_tokens":15341352,"total_runtime_s":416112,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.189296504626746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-377fac1560bdaff4e09d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-fa3d4b7ff17b05ed7e112fbe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4300608344770858,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1364,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[681];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1364,"metadata":{"avgScore":0.4300608344770858,"avg_score":0.4300608344770858,"avg_score_percent":43.00608344770858,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1508.33,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1571377795,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5621171,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":135,"runtime":267844,"source_flags":[],"source_harness":"claude_code","source_row_index":681,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1508.33,"totalDurationS":267844,"totalInputTokens":1571377795,"totalOutputTokens":5621171,"total_cost_usd":1508.33,"total_input_tokens":1571377795,"total_output_tokens":5621171,"total_runtime_s":267844,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.00608344770858,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-26907115f2430cfac841","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f800508b3a85c417e03a9ea7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.08,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":672,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[335];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":672,"metadata":{"avgScore":0.08,"avg_score":0.08,"avg_score_percent":8.0,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":509.28,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":144741842,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":380341,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":23712,"source_flags":[],"source_harness":"claude_code","source_row_index":335,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":509.28,"totalDurationS":23712,"totalInputTokens":144741842,"totalOutputTokens":380341,"total_cost_usd":509.28,"total_input_tokens":144741842,"total_output_tokens":380341,"total_runtime_s":23712,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eefb3def96aa0693413e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f2f4e5aca0db5b83f0c5dbf3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.41885525446026206,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":980,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[489];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":980,"metadata":{"avgScore":0.41885525446026206,"avg_score":0.41885525446026206,"avg_score_percent":41.88552544602621,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1203.45,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":396448337,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":7507408,"passRate":0.24,"pass_rate":0.24,"pass_rate_percent":24.0,"passes":12,"runs":50,"runtime":174938,"source_flags":[],"source_harness":"claude_code","source_row_index":489,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":1203.45,"totalDurationS":174938,"totalInputTokens":396448337,"totalOutputTokens":7507408,"total_cost_usd":1203.45,"total_input_tokens":396448337,"total_output_tokens":7507408,"total_runtime_s":174938,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.88552544602621,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e0e86319f0ed30f61756","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ef2afcbf5b4dec6583c95715","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.598098576643749,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":428,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[213];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":428,"metadata":{"avgScore":0.598098576643749,"avg_score":0.598098576643749,"avg_score_percent":59.8098576643749,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":397673605,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":2241277,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":76504,"source_flags":[],"source_harness":"claude_code","source_row_index":213,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":446,"totalDurationS":76504,"totalInputTokens":397673605,"totalOutputTokens":2241277,"total_cost_usd":446,"total_input_tokens":397673605,"total_output_tokens":2241277,"total_runtime_s":76504,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":59.8098576643749,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c3347ca2f461ee5b73a4","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-eb7345bcb32cba01115e24d2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1083130564079132,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":664,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[331];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":664,"metadata":{"avgScore":0.1083130564079132,"avg_score":0.1083130564079132,"avg_score_percent":10.831305640791319,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":117.67,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":225487387,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":811747,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":8,"runtime":30318,"source_flags":[],"source_harness":"claude_code","source_row_index":331,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":8,"tokens":null,"totalCostUsd":117.67,"totalDurationS":30318,"totalInputTokens":225487387,"totalOutputTokens":811747,"total_cost_usd":117.67,"total_input_tokens":225487387,"total_output_tokens":811747,"total_runtime_s":30318,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.831305640791319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d22e5a0be1ff0ad5e2a7","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e89379d9c5e013413f70423d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.4067685680771121,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1382,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[690];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1382,"metadata":{"avgScore":0.4067685680771121,"avg_score":0.4067685680771121,"avg_score_percent":40.67685680771121,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1328.51,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":307300337,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":3412910,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":136,"runtime":466272,"source_flags":[],"source_harness":"claude_code","source_row_index":690,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1328.51,"totalDurationS":466272,"totalInputTokens":307300337,"totalOutputTokens":3412910,"total_cost_usd":1328.51,"total_input_tokens":307300337,"total_output_tokens":3412910,"total_runtime_s":466272,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.67685680771121,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-591b1c5b26842505d385","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e527d501ae73a0b6ed4ab6c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.4089984699607975,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":782,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[390];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":782,"metadata":{"avgScore":0.4089984699607975,"avg_score":0.4089984699607975,"avg_score_percent":40.899846996079745,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":566.31,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":112447742,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1784448,"passRate":0.17142857142857143,"pass_rate":0.17142857142857143,"pass_rate_percent":17.142857142857142,"passes":18,"runs":102,"runtime":324041,"source_flags":[],"source_harness":"claude_code","source_row_index":390,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":566.31,"totalDurationS":324041,"totalInputTokens":112447742,"totalOutputTokens":1784448,"total_cost_usd":566.31,"total_input_tokens":112447742,"total_output_tokens":1784448,"total_runtime_s":324041,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.899846996079745,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7d9e1d96aeb4185a22bf","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e1f543cbc22d0f7882bff161","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6273164650789494,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1240,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[619];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1240,"metadata":{"avgScore":0.6273164650789494,"avg_score":0.6273164650789494,"avg_score_percent":62.73164650789494,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":683.28,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":201294188,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5981020,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":66,"runtime":151680,"source_flags":[],"source_harness":"claude_code","source_row_index":619,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":683.28,"totalDurationS":151680,"totalInputTokens":201294188,"totalOutputTokens":5981020,"total_cost_usd":683.28,"total_input_tokens":201294188,"total_output_tokens":5981020,"total_runtime_s":151680,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":62.73164650789494,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-68ad083d41b85eb2d992","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e1145ce6bfa874c11d2533f3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40677645580102734,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":534,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[266];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":534,"metadata":{"avgScore":0.40677645580102734,"avg_score":0.40677645580102734,"avg_score_percent":40.677645580102734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1626,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1796865182,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":6432918,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":143,"runtime":298162,"source_flags":[],"source_harness":"claude_code","source_row_index":266,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1626,"totalDurationS":298162,"totalInputTokens":1796865182,"totalOutputTokens":6432918,"total_cost_usd":1626,"total_input_tokens":1796865182,"total_output_tokens":6432918,"total_runtime_s":298162,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.677645580102734,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8218083c1e7392228edc","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-db0552e1b08b92c7f03f214d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.38265263820134726,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":166,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[82];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":166,"metadata":{"avgScore":0.38265263820134726,"avg_score":0.38265263820134726,"avg_score_percent":38.26526382013473,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":855.55,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":300355769,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5612384,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":55,"runtime":158175,"source_flags":[],"source_harness":"claude_code","source_row_index":82,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":855.55,"totalDurationS":158175,"totalInputTokens":300355769,"totalOutputTokens":5612384,"total_cost_usd":855.55,"total_input_tokens":300355769,"total_output_tokens":5612384,"total_runtime_s":158175,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.26526382013473,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7908f159c80713942942","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d5c3797ab42eeba8a066e6c0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.0856096444480286,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1150,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[574];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1150,"metadata":{"avgScore":0.0856096444480286,"avg_score":0.0856096444480286,"avg_score_percent":8.56096444480286,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":580.14,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":726419627,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1662333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":26,"runtime":135036,"source_flags":[],"source_harness":"claude_code","source_row_index":574,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":580.14,"totalDurationS":135036,"totalInputTokens":726419627,"totalOutputTokens":1662333,"total_cost_usd":580.14,"total_input_tokens":726419627,"total_output_tokens":1662333,"total_runtime_s":135036,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":8.56096444480286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8320b705f6da9a1b0dc7","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d21c292c163a27d940e884cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6273164650789494,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":404,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[201];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":404,"metadata":{"avgScore":0.6273164650789494,"avg_score":0.6273164650789494,"avg_score_percent":62.73164650789494,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":683.28,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":201294188,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5981020,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":66,"runtime":151680,"source_flags":[],"source_harness":"claude_code","source_row_index":201,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":683.28,"totalDurationS":151680,"totalInputTokens":201294188,"totalOutputTokens":5981020,"total_cost_usd":683.28,"total_input_tokens":201294188,"total_output_tokens":5981020,"total_runtime_s":151680,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":62.73164650789494,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b651754869985c760c2c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-cc3efd05f55bff8f1146ceaa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.18756758749923322,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":638,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[318];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":638,"metadata":{"avgScore":0.18756758749923322,"avg_score":0.18756758749923322,"avg_score_percent":18.75675874992332,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":203.72,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":74904373,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1166124,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":103208,"source_flags":[],"source_harness":"claude_code","source_row_index":318,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":203.72,"totalDurationS":103208,"totalInputTokens":74904373,"totalOutputTokens":1166124,"total_cost_usd":203.72,"total_input_tokens":74904373,"total_output_tokens":1166124,"total_runtime_s":103208,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.75675874992332,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f5554851f0a02621d69e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c96a094e22e671eaa9c0aa57","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2873379749190263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":200,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[99];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":200,"metadata":{"avgScore":0.2873379749190263,"avg_score":0.2873379749190263,"avg_score_percent":28.733797491902628,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":485.88,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":118264612,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1117633,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":54,"runtime":171502,"source_flags":[],"source_harness":"claude_code","source_row_index":99,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":485.88,"totalDurationS":171502,"totalInputTokens":118264612,"totalOutputTokens":1117633,"total_cost_usd":485.88,"total_input_tokens":118264612,"total_output_tokens":1117633,"total_runtime_s":171502,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.733797491902628,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2bf849c019cc8fbe175c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ae8ccd9162636bbea853963e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4234878390395742,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":50,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[24];split=full;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":50,"metadata":{"avgScore":0.4234878390395742,"avg_score":0.4234878390395742,"avg_score_percent":42.34878390395742,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2815,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1209502455,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":16507476,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":151,"runtime":519320,"source_flags":[],"source_harness":"claude_code","source_row_index":24,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":2815,"totalDurationS":519320,"totalInputTokens":1209502455,"totalOutputTokens":16507476,"total_cost_usd":2815,"total_input_tokens":1209502455,"total_output_tokens":16507476,"total_runtime_s":519320,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.34878390395742,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e10eb6a0c0ce69973526","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a4394b9d8f99c66bbd4ffefa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10431719138556748,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":298,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[148];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":298,"metadata":{"avgScore":0.10431719138556748,"avg_score":0.10431719138556748,"avg_score_percent":10.431719138556748,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1395.12,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":739811872,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5918841,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":38,"runtime":253733,"source_flags":[],"source_harness":"claude_code","source_row_index":148,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":1395.12,"totalDurationS":253733,"totalInputTokens":739811872,"totalOutputTokens":5918841,"total_cost_usd":1395.12,"total_input_tokens":739811872,"total_output_tokens":5918841,"total_runtime_s":253733,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.431719138556748,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-adaddc939beb96afaa2f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a4353cba2c92cade86963004","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.38312084275574215,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":76,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[37];split=full;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":76,"metadata":{"avgScore":0.38312084275574215,"avg_score":0.38312084275574215,"avg_score_percent":38.312084275574215,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1837.79,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":452042179,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":3793251,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":141,"runtime":489984,"source_flags":[],"source_harness":"claude_code","source_row_index":37,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1837.79,"totalDurationS":489984,"totalInputTokens":452042179,"totalOutputTokens":3793251,"total_cost_usd":1837.79,"total_input_tokens":452042179,"total_output_tokens":3793251,"total_runtime_s":489984,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.312084275574215,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5059301f37d8ed4b1a47","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a3dcf10d9e289695c08b888b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.6172996937063634,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":412,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[205];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":412,"metadata":{"avgScore":0.6172996937063634,"avg_score":0.6172996937063634,"avg_score_percent":61.72996937063634,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":805.69,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":190131175,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1966191,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":156896,"source_flags":[],"source_harness":"claude_code","source_row_index":205,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":805.69,"totalDurationS":156896,"totalInputTokens":190131175,"totalOutputTokens":1966191,"total_cost_usd":805.69,"total_input_tokens":190131175,"total_output_tokens":1966191,"total_runtime_s":156896,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":61.72996937063634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2da3ccf04d11b75c8c9b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a2580adb1510b85cfb9e1845","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4694146393591237,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1346,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[672];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1346,"metadata":{"avgScore":0.4694146393591237,"avg_score":0.4694146393591237,"avg_score_percent":46.94146393591237,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3696.61,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1465162311,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":20372367,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":139,"runtime":543814,"source_flags":[],"source_harness":"claude_code","source_row_index":672,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":3696.61,"totalDurationS":543814,"totalInputTokens":1465162311,"totalOutputTokens":20372367,"total_cost_usd":3696.61,"total_input_tokens":1465162311,"total_output_tokens":20372367,"total_runtime_s":543814,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.94146393591237,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-306b2e00af0108706fd9","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-9e214b6110d22d84bb80b783","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44189296504626746,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":888,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[443];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":888,"metadata":{"avgScore":0.44189296504626746,"avg_score":0.44189296504626746,"avg_score_percent":44.189296504626746,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2611.28,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1134598082,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":15341352,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":140,"runtime":416112,"source_flags":[],"source_harness":"claude_code","source_row_index":443,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":2611.28,"totalDurationS":416112,"totalInputTokens":1134598082,"totalOutputTokens":15341352,"total_cost_usd":2611.28,"total_input_tokens":1134598082,"total_output_tokens":15341352,"total_runtime_s":416112,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.189296504626746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-38c02f6c4e4b1aa872c7","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-9c6e0574619ad7a163e4e9e6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4514857743894153,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":510,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[254];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":510,"metadata":{"avgScore":0.4514857743894153,"avg_score":0.4514857743894153,"avg_score_percent":45.14857743894153,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3985,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1692408720,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":22067608,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":150,"runtime":645508,"source_flags":[],"source_harness":"claude_code","source_row_index":254,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":3985,"totalDurationS":645508,"totalInputTokens":1692408720,"totalOutputTokens":22067608,"total_cost_usd":3985,"total_input_tokens":1692408720,"total_output_tokens":22067608,"total_runtime_s":645508,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.14857743894153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e09d63a584741b3c7d3b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-99a18c42b6f8a7369e700785","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6396884267777608,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":402,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[200];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":402,"metadata":{"avgScore":0.6396884267777608,"avg_score":0.6396884267777608,"avg_score_percent":63.96884267777608,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1105.97,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":310513347,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":8274090,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":66,"runtime":178788,"source_flags":[],"source_harness":"claude_code","source_row_index":200,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":1105.97,"totalDurationS":178788,"totalInputTokens":310513347,"totalOutputTokens":8274090,"total_cost_usd":1105.97,"total_input_tokens":310513347,"total_output_tokens":8274090,"total_runtime_s":178788,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":63.96884267777608,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d05648f8b31e40d5dbcb","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-9335da1296291a850210f2f7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4694146393591237,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":874,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[436];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":874,"metadata":{"avgScore":0.4694146393591237,"avg_score":0.4694146393591237,"avg_score_percent":46.94146393591237,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3696.61,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1465162311,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":20372367,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":139,"runtime":543814,"source_flags":[],"source_harness":"claude_code","source_row_index":436,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":3696.61,"totalDurationS":543814,"totalInputTokens":1465162311,"totalOutputTokens":20372367,"total_cost_usd":3696.61,"total_input_tokens":1465162311,"total_output_tokens":20372367,"total_runtime_s":543814,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.94146393591237,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4e8a2362420d8aa293bd","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8fb41b7e0c4c451bac83bef7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.42142958861655383,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":776,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[387];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":776,"metadata":{"avgScore":0.42142958861655383,"avg_score":0.42142958861655383,"avg_score_percent":42.14295886165538,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1713.22,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":640997605,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":11872346,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":21,"runs":104,"runtime":304614,"source_flags":[],"source_harness":"claude_code","source_row_index":387,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":1713.22,"totalDurationS":304614,"totalInputTokens":640997605,"totalOutputTokens":11872346,"total_cost_usd":1713.22,"total_input_tokens":640997605,"total_output_tokens":11872346,"total_runtime_s":304614,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.14295886165538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a9a01ee934c078dbf8c6","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-899d81e62ee5606f35909d7c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3793836589504913,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1004,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[501];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1004,"metadata":{"avgScore":0.3793836589504913,"avg_score":0.3793836589504913,"avg_score_percent":37.93836589504913,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":777.77,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":255949496,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5279930,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":10,"runs":50,"runtime":115789,"source_flags":[],"source_harness":"claude_code","source_row_index":501,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":777.77,"totalDurationS":115789,"totalInputTokens":255949496,"totalOutputTokens":5279930,"total_cost_usd":777.77,"total_input_tokens":255949496,"total_output_tokens":5279930,"total_runtime_s":115789,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.93836589504913,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9af31ef10e0e3ff44b44","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8322d64c35cec94e9039eba5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.069505480733747,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":322,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[160];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":322,"metadata":{"avgScore":0.069505480733747,"avg_score":0.069505480733747,"avg_score_percent":6.9505480733747005,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":734.85,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":841102097,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":2207676,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":32,"runtime":155287,"source_flags":[],"source_harness":"claude_code","source_row_index":160,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":734.85,"totalDurationS":155287,"totalInputTokens":841102097,"totalOutputTokens":2207676,"total_cost_usd":734.85,"total_input_tokens":841102097,"total_output_tokens":2207676,"total_runtime_s":155287,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.9505480733747005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fabc3c87725feec871c5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-6d252c1e6c3a94493bbf8537","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1440505369585742,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":266,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[132];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":266,"metadata":{"avgScore":0.1440505369585742,"avg_score":0.1440505369585742,"avg_score_percent":14.40505369585742,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1805.3,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":961017525,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":7274340,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":37,"runtime":305062,"source_flags":[],"source_harness":"claude_code","source_row_index":132,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":1805.3,"totalDurationS":305062,"totalInputTokens":961017525,"totalOutputTokens":7274340,"total_cost_usd":1805.3,"total_input_tokens":961017525,"total_output_tokens":7274340,"total_runtime_s":305062,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.40505369585742,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f4e05d1ca72631891f7c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-615eeb0233cb19bbb47ca78b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.052413157894736835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":330,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[164];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":330,"metadata":{"avgScore":0.052413157894736835,"avg_score":0.052413157894736835,"avg_score_percent":5.241315789473684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":601.73,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":157559357,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":870700,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":27,"runtime":186747,"source_flags":[],"source_harness":"claude_code","source_row_index":164,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":27,"tokens":null,"totalCostUsd":601.73,"totalDurationS":186747,"totalInputTokens":157559357,"totalOutputTokens":870700,"total_cost_usd":601.73,"total_input_tokens":157559357,"total_output_tokens":870700,"total_runtime_s":186747,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":5.241315789473684,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d409de710c16f69c253c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-60cb28e26731d8df323b6feb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.39611287961350156,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":154,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[76];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":154,"metadata":{"avgScore":0.39611287961350156,"avg_score":0.39611287961350156,"avg_score_percent":39.61128796135016,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":568.52,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":698108254,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":2255196,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":53,"runtime":94079,"source_flags":[],"source_harness":"claude_code","source_row_index":76,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":568.52,"totalDurationS":94079,"totalInputTokens":698108254,"totalOutputTokens":2255196,"total_cost_usd":568.52,"total_input_tokens":698108254,"total_output_tokens":2255196,"total_runtime_s":94079,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.61128796135016,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a53a21b9639ea4e7fc3e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5d03ca437284a8da918d17cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.6172996937063634,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1248,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[623];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1248,"metadata":{"avgScore":0.6172996937063634,"avg_score":0.6172996937063634,"avg_score_percent":61.72996937063634,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":805.69,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":190131175,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1966191,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":156896,"source_flags":[],"source_harness":"claude_code","source_row_index":623,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":805.69,"totalDurationS":156896,"totalInputTokens":190131175,"totalOutputTokens":1966191,"total_cost_usd":805.69,"total_input_tokens":190131175,"total_output_tokens":1966191,"total_runtime_s":156896,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":61.72996937063634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c375c31b4ecabbbcde1f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-58204a5f86cac628ca3dfb96","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.598098576643749,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1264,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[631];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1264,"metadata":{"avgScore":0.598098576643749,"avg_score":0.598098576643749,"avg_score_percent":59.8098576643749,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":397673605,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":2241277,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":76504,"source_flags":[],"source_harness":"claude_code","source_row_index":631,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":446,"totalDurationS":76504,"totalInputTokens":397673605,"totalOutputTokens":2241277,"total_cost_usd":446,"total_input_tokens":397673605,"total_output_tokens":2241277,"total_runtime_s":76504,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":59.8098576643749,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-36467ec6de85fde64273","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-546209e23c10f13bddc5ba2a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40677645580102734,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":62,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[30];split=full;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":62,"metadata":{"avgScore":0.40677645580102734,"avg_score":0.40677645580102734,"avg_score_percent":40.677645580102734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1626,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1796865182,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":6432918,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":143,"runtime":298162,"source_flags":[],"source_harness":"claude_code","source_row_index":30,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1626,"totalDurationS":298162,"totalInputTokens":1796865182,"totalOutputTokens":6432918,"total_cost_usd":1626,"total_input_tokens":1796865182,"total_output_tokens":6432918,"total_runtime_s":298162,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.677645580102734,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-30591eb721567bc7259a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-520eb707155fa26c59056fbb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.09925073212365117,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1136,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[567];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1136,"metadata":{"avgScore":0.09925073212365117,"avg_score":0.09925073212365117,"avg_score_percent":9.925073212365117,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1101.41,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":706933041,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":4977921,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":30,"runtime":173011,"source_flags":[],"source_harness":"claude_code","source_row_index":567,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":1101.41,"totalDurationS":173011,"totalInputTokens":706933041,"totalOutputTokens":4977921,"total_cost_usd":1101.41,"total_input_tokens":706933041,"total_output_tokens":4977921,"total_runtime_s":173011,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.925073212365117,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1c8ee0e2d85c0b19e037","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4a990bb8c7a5756b05a3bcd2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3933536738539414,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":994,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[496];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":994,"metadata":{"avgScore":0.3933536738539414,"avg_score":0.3933536738539414,"avg_score_percent":39.335367385394136,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":516.84,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":544268438,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1875055,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":10,"runs":49,"runtime":76526,"source_flags":[],"source_harness":"claude_code","source_row_index":496,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":516.84,"totalDurationS":76526,"totalInputTokens":544268438,"totalOutputTokens":1875055,"total_cost_usd":516.84,"total_input_tokens":544268438,"total_output_tokens":1875055,"total_runtime_s":76526,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.335367385394136,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92c9277bdef663604d77","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4a91e8593f90936acb639ff4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4234878390395742,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":522,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[260];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":522,"metadata":{"avgScore":0.4234878390395742,"avg_score":0.4234878390395742,"avg_score_percent":42.34878390395742,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2815,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1209502455,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":16507476,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":151,"runtime":519320,"source_flags":[],"source_harness":"claude_code","source_row_index":260,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":2815,"totalDurationS":519320,"totalInputTokens":1209502455,"totalOutputTokens":16507476,"total_cost_usd":2815,"total_input_tokens":1209502455,"total_output_tokens":16507476,"total_runtime_s":519320,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.34878390395742,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b251248ff7ac67bdfce1","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-49e76cb1785db950f66506ee","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6396884267777608,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1238,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[618];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1238,"metadata":{"avgScore":0.6396884267777608,"avg_score":0.6396884267777608,"avg_score_percent":63.96884267777608,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1105.97,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":310513347,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":8274090,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":66,"runtime":178788,"source_flags":[],"source_harness":"claude_code","source_row_index":618,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":1105.97,"totalDurationS":178788,"totalInputTokens":310513347,"totalOutputTokens":8274090,"total_cost_usd":1105.97,"total_input_tokens":310513347,"total_output_tokens":8274090,"total_runtime_s":178788,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":63.96884267777608,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b69664d0abc35765207a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3f6fb2c137523da5efd690ee","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.4067685680771121,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":910,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[454];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":910,"metadata":{"avgScore":0.4067685680771121,"avg_score":0.4067685680771121,"avg_score_percent":40.67685680771121,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1328.51,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":307300337,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":3412910,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":136,"runtime":466272,"source_flags":[],"source_harness":"claude_code","source_row_index":454,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1328.51,"totalDurationS":466272,"totalInputTokens":307300337,"totalOutputTokens":3412910,"total_cost_usd":1328.51,"total_input_tokens":307300337,"total_output_tokens":3412910,"total_runtime_s":466272,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.67685680771121,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1ef842ed00ffc8654cc2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3caad62dcf2ab1db17d2d0b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4406278923915453,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":760,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[379];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":760,"metadata":{"avgScore":0.4406278923915453,"avg_score":0.4406278923915453,"avg_score_percent":44.06278923915453,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1953.22,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":642239869,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":15272399,"passRate":0.2571428571428571,"pass_rate":0.2571428571428571,"pass_rate_percent":25.71428571428571,"passes":27,"runs":104,"runtime":399408,"source_flags":[],"source_harness":"claude_code","source_row_index":379,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":1953.22,"totalDurationS":399408,"totalInputTokens":642239869,"totalOutputTokens":15272399,"total_cost_usd":1953.22,"total_input_tokens":642239869,"total_output_tokens":15272399,"total_runtime_s":399408,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.06278923915453,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a09ce1e31f2c89f6d0fa","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-31ac45ddbfe11b4ddda7caa5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.38312084275574215,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":548,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[273];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":548,"metadata":{"avgScore":0.38312084275574215,"avg_score":0.38312084275574215,"avg_score_percent":38.312084275574215,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1837.79,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":452042179,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":3793251,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":141,"runtime":489984,"source_flags":[],"source_harness":"claude_code","source_row_index":273,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1837.79,"totalDurationS":489984,"totalInputTokens":452042179,"totalOutputTokens":3793251,"total_cost_usd":1837.79,"total_input_tokens":452042179,"total_output_tokens":3793251,"total_runtime_s":489984,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.312084275574215,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0af77329830872c19772","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-317041e05913eb7147feb601","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4300608344770858,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":892,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[445];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":892,"metadata":{"avgScore":0.4300608344770858,"avg_score":0.4300608344770858,"avg_score_percent":43.00608344770858,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1508.33,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1571377795,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":5621171,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":135,"runtime":267844,"source_flags":[],"source_harness":"claude_code","source_row_index":445,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1508.33,"totalDurationS":267844,"totalInputTokens":1571377795,"totalOutputTokens":5621171,"total_cost_usd":1508.33,"total_input_tokens":1571377795,"total_output_tokens":5621171,"total_runtime_s":267844,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.00608344770858,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c27dffdcbfc14ec0298e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2f466881d8321e221ba1a6b0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.06638999999999999,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1164,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[581];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1164,"metadata":{"avgScore":0.06638999999999999,"avg_score":0.06638999999999999,"avg_score_percent":6.638999999999999,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":158.25,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":32912598,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":542905,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":25,"runtime":174170,"source_flags":[],"source_harness":"claude_code","source_row_index":581,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":25,"tokens":null,"totalCostUsd":158.25,"totalDurationS":174170,"totalInputTokens":32912598,"totalOutputTokens":542905,"total_cost_usd":158.25,"total_input_tokens":32912598,"total_output_tokens":542905,"total_runtime_s":174170,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.638999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a3a41772d54060e23593","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-20df2a57c0c8d2da49bb954b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4514857743894153,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":38,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[18];split=full;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":38,"metadata":{"avgScore":0.4514857743894153,"avg_score":0.4514857743894153,"avg_score_percent":45.14857743894153,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3985,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1692408720,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":22067608,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":150,"runtime":645508,"source_flags":[],"source_harness":"claude_code","source_row_index":18,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":3985,"totalDurationS":645508,"totalInputTokens":1692408720,"totalOutputTokens":22067608,"total_cost_usd":3985,"total_input_tokens":1692408720,"total_output_tokens":22067608,"total_runtime_s":645508,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.14857743894153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-94d13975ff98860287ed","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1e85b252dab80edab378bc6c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4460039363451024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":756,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[377];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":756,"metadata":{"avgScore":0.4460039363451024,"avg_score":0.4460039363451024,"avg_score_percent":44.60039363451024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1015.69,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1098101909,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":4222336,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":25,"runs":102,"runtime":211355,"source_flags":[],"source_harness":"claude_code","source_row_index":377,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":1015.69,"totalDurationS":211355,"totalInputTokens":1098101909,"totalOutputTokens":4222336,"total_cost_usd":1015.69,"total_input_tokens":1098101909,"total_output_tokens":4222336,"total_runtime_s":211355,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.60039363451024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5ba074117de97948bf67","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-16ef764703fbf32568c3d319","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2984717724109289,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1034,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[516];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1034,"metadata":{"avgScore":0.2984717724109289,"avg_score":0.2984717724109289,"avg_score_percent":29.847177241092893,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.82,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":97287341,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1049930,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":50,"runtime":159935,"source_flags":[],"source_harness":"claude_code","source_row_index":516,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":415.82,"totalDurationS":159935,"totalInputTokens":97287341,"totalOutputTokens":1049930,"total_cost_usd":415.82,"total_input_tokens":97287341,"total_output_tokens":1049930,"total_runtime_s":159935,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.847177241092893,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-566b435ce68e8d2756a1","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-125e62636a10b0af78eb71a2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22167032341406182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":618,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[308];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":618,"metadata":{"avgScore":0.22167032341406182,"avg_score":0.22167032341406182,"avg_score_percent":22.16703234140618,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":288.39,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":227246409,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":1695241,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":101694,"source_flags":[],"source_harness":"claude_code","source_row_index":308,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":288.39,"totalDurationS":101694,"totalInputTokens":227246409,"totalOutputTokens":1695241,"total_cost_usd":288.39,"total_input_tokens":227246409,"total_output_tokens":1695241,"total_runtime_s":101694,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.16703234140618,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f92ed22808018f5bdacf","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0f737ad0795a98db4eafcbb1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.16051822822903797,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1094,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[546];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1094,"metadata":{"avgScore":0.16051822822903797,"avg_score":0.16051822822903797,"avg_score_percent":16.051822822903798,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1425.24,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":858187522,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":6121545,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":29,"runtime":225064,"source_flags":[],"source_harness":"claude_code","source_row_index":546,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":1425.24,"totalDurationS":225064,"totalInputTokens":858187522,"totalOutputTokens":6121545,"total_cost_usd":1425.24,"total_input_tokens":858187522,"total_output_tokens":6121545,"total_runtime_s":225064,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.051822822903798,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dce97d54d2a14db2af73","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-04fb8c1f2aa0ea40fd594cb2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4313229586002382,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72d37ce12f60a534be336a3519e5905d7f50699a","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":138,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[68];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":138,"metadata":{"avgScore":0.4313229586002382,"avg_score":0.4313229586002382,"avg_score_percent":43.132295860023824,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1323.79,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":524284122,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-4-8","output_tokens":8213812,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":55,"runtime":214602,"source_flags":[],"source_harness":"claude_code","source_row_index":68,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":1323.79,"totalDurationS":214602,"totalInputTokens":524284122,"totalOutputTokens":8213812,"total_cost_usd":1323.79,"total_input_tokens":524284122,"total_output_tokens":8213812,"total_runtime_s":214602,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.132295860023824,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e91b62a4867e2b199950","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ffb14ef9a96458679d7d6771","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":891,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[445];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":891,"metadata":{"avgScore":0.4300608344770858,"avg_score":0.4300608344770858,"avg_score_percent":43.00608344770858,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1508.33,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1571377795,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5621171,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":135,"runtime":267844,"source_flags":[],"source_harness":"claude_code","source_row_index":445,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1508.33,"totalDurationS":267844,"totalInputTokens":1571377795,"totalOutputTokens":5621171,"total_cost_usd":1508.33,"total_input_tokens":1571377795,"total_output_tokens":5621171,"total_runtime_s":267844,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fe7551b74ae3876b2819","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-fc735ea33ea3cd999fa1a6ac","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3582089552238806,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1239,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[619];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1239,"metadata":{"avgScore":0.6273164650789494,"avg_score":0.6273164650789494,"avg_score_percent":62.73164650789494,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":683.28,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":201294188,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5981020,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":66,"runtime":151680,"source_flags":[],"source_harness":"claude_code","source_row_index":619,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":683.28,"totalDurationS":151680,"totalInputTokens":201294188,"totalOutputTokens":5981020,"total_cost_usd":683.28,"total_input_tokens":201294188,"total_output_tokens":5981020,"total_runtime_s":151680,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.82089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f8af3793b079ae73aa94","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f68953aedee2c40676819644","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.16447368421052633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":75,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[37];split=full;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":75,"metadata":{"avgScore":0.38312084275574215,"avg_score":0.38312084275574215,"avg_score_percent":38.312084275574215,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1837.79,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":452042179,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":3793251,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":141,"runtime":489984,"source_flags":[],"source_harness":"claude_code","source_row_index":37,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1837.79,"totalDurationS":489984,"totalInputTokens":452042179,"totalOutputTokens":3793251,"total_cost_usd":1837.79,"total_input_tokens":452042179,"total_output_tokens":3793251,"total_runtime_s":489984,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.447368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e01a03f9ed169e85be06","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f1fe44666370853c84d854f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.373134328358209,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1263,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[631];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1263,"metadata":{"avgScore":0.598098576643749,"avg_score":0.598098576643749,"avg_score_percent":59.8098576643749,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":397673605,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":2241277,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":76504,"source_flags":[],"source_harness":"claude_code","source_row_index":631,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":446,"totalDurationS":76504,"totalInputTokens":397673605,"totalOutputTokens":2241277,"total_cost_usd":446,"total_input_tokens":397673605,"total_output_tokens":2241277,"total_runtime_s":76504,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.3134328358209,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b2c3ab67f69773613f76","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e706c1c7f24f3bc33be1edb1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":165,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[82];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":165,"metadata":{"avgScore":0.38265263820134726,"avg_score":0.38265263820134726,"avg_score_percent":38.26526382013473,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":855.55,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":300355769,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5612384,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":55,"runtime":158175,"source_flags":[],"source_harness":"claude_code","source_row_index":82,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":855.55,"totalDurationS":158175,"totalInputTokens":300355769,"totalOutputTokens":5612384,"total_cost_usd":855.55,"total_input_tokens":300355769,"total_output_tokens":5612384,"total_runtime_s":158175,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b11200666d3cb6b82ac9","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-df949fa563bc4e10e761201e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2907801418439716,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1345,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[672];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1345,"metadata":{"avgScore":0.4694146393591237,"avg_score":0.4694146393591237,"avg_score_percent":46.94146393591237,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3696.61,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1465162311,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":20372367,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":139,"runtime":543814,"source_flags":[],"source_harness":"claude_code","source_row_index":672,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":3696.61,"totalDurationS":543814,"totalInputTokens":1465162311,"totalOutputTokens":20372367,"total_cost_usd":3696.61,"total_input_tokens":1465162311,"total_output_tokens":20372367,"total_runtime_s":543814,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.078014184397162,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6a056ff9f1ee13f9d6cc","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-dc7e35f68834bc985be07af0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":979,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[489];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":979,"metadata":{"avgScore":0.41885525446026206,"avg_score":0.41885525446026206,"avg_score_percent":41.88552544602621,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1203.45,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":396448337,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":7507408,"passRate":0.24,"pass_rate":0.24,"pass_rate_percent":24.0,"passes":12,"runs":50,"runtime":174938,"source_flags":[],"source_harness":"claude_code","source_row_index":489,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":1203.45,"totalDurationS":174938,"totalInputTokens":396448337,"totalOutputTokens":7507408,"total_cost_usd":1203.45,"total_input_tokens":396448337,"total_output_tokens":7507408,"total_runtime_s":174938,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eda6e9bc9ab12550ee27","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d728065782f33ce56899afa1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1135,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[567];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1135,"metadata":{"avgScore":0.09925073212365117,"avg_score":0.09925073212365117,"avg_score_percent":9.925073212365117,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1101.41,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":706933041,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":4977921,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":30,"runtime":173011,"source_flags":[],"source_harness":"claude_code","source_row_index":567,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":1101.41,"totalDurationS":173011,"totalInputTokens":706933041,"totalOutputTokens":4977921,"total_cost_usd":1101.41,"total_input_tokens":706933041,"total_output_tokens":4977921,"total_runtime_s":173011,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8197b5bb9ae059fcf3b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d5b70b9805fab1c3023373e0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23809523809523808,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":755,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[377];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":755,"metadata":{"avgScore":0.4460039363451024,"avg_score":0.4460039363451024,"avg_score_percent":44.60039363451024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1015.69,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1098101909,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":4222336,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":25,"runs":102,"runtime":211355,"source_flags":[],"source_harness":"claude_code","source_row_index":377,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":1015.69,"totalDurationS":211355,"totalInputTokens":1098101909,"totalOutputTokens":4222336,"total_cost_usd":1015.69,"total_input_tokens":1098101909,"total_output_tokens":4222336,"total_runtime_s":211355,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.809523809523807,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c3e79e256c5cc67a4940","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d2a2a557c23a3790fb3d7489","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1363,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[681];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1363,"metadata":{"avgScore":0.4300608344770858,"avg_score":0.4300608344770858,"avg_score_percent":43.00608344770858,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1508.33,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1571377795,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5621171,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":135,"runtime":267844,"source_flags":[],"source_harness":"claude_code","source_row_index":681,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1508.33,"totalDurationS":267844,"totalInputTokens":1571377795,"totalOutputTokens":5621171,"total_cost_usd":1508.33,"total_input_tokens":1571377795,"total_output_tokens":5621171,"total_runtime_s":267844,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a52aa6d42a2ccd6e18d4","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c2e3be55d8fed62c7a7b0101","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":153,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[76];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":153,"metadata":{"avgScore":0.39611287961350156,"avg_score":0.39611287961350156,"avg_score_percent":39.61128796135016,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":568.52,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":698108254,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":2255196,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":53,"runtime":94079,"source_flags":[],"source_harness":"claude_code","source_row_index":76,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":568.52,"totalDurationS":94079,"totalInputTokens":698108254,"totalOutputTokens":2255196,"total_cost_usd":568.52,"total_input_tokens":698108254,"total_output_tokens":2255196,"total_runtime_s":94079,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7fee32969ab57432af7c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c14fcbd230ffa719469eb611","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":265,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[132];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":265,"metadata":{"avgScore":0.1440505369585742,"avg_score":0.1440505369585742,"avg_score_percent":14.40505369585742,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1805.3,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":961017525,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":7274340,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":37,"runtime":305062,"source_flags":[],"source_harness":"claude_code","source_row_index":132,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":1805.3,"totalDurationS":305062,"totalInputTokens":961017525,"totalOutputTokens":7274340,"total_cost_usd":1805.3,"total_input_tokens":961017525,"total_output_tokens":7274340,"total_runtime_s":305062,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-65dbb1b0309691c49351","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c0fd5e235cb823a5a0a5b59a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":321,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[160];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":321,"metadata":{"avgScore":0.069505480733747,"avg_score":0.069505480733747,"avg_score_percent":6.9505480733747005,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":734.85,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":841102097,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":2207676,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":32,"runtime":155287,"source_flags":[],"source_harness":"claude_code","source_row_index":160,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":734.85,"totalDurationS":155287,"totalInputTokens":841102097,"totalOutputTokens":2207676,"total_cost_usd":734.85,"total_input_tokens":841102097,"total_output_tokens":2207676,"total_runtime_s":155287,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ee8ee15c4d91eaa356c2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-bd0e984dc37b5aa20d38d8d0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1093,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[546];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1093,"metadata":{"avgScore":0.16051822822903797,"avg_score":0.16051822822903797,"avg_score_percent":16.051822822903798,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1425.24,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":858187522,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":6121545,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":29,"runtime":225064,"source_flags":[],"source_harness":"claude_code","source_row_index":546,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":1425.24,"totalDurationS":225064,"totalInputTokens":858187522,"totalOutputTokens":6121545,"total_cost_usd":1425.24,"total_input_tokens":858187522,"total_output_tokens":6121545,"total_runtime_s":225064,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-727b28b0ecc3656b3c8f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-bc3cd3e1a34eaf4d28c590ba","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.17142857142857143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":781,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[390];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":781,"metadata":{"avgScore":0.4089984699607975,"avg_score":0.4089984699607975,"avg_score_percent":40.899846996079745,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":566.31,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":112447742,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1784448,"passRate":0.17142857142857143,"pass_rate":0.17142857142857143,"pass_rate_percent":17.142857142857142,"passes":18,"runs":102,"runtime":324041,"source_flags":[],"source_harness":"claude_code","source_row_index":390,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":566.31,"totalDurationS":324041,"totalInputTokens":112447742,"totalOutputTokens":1784448,"total_cost_usd":566.31,"total_input_tokens":112447742,"total_output_tokens":1784448,"total_runtime_s":324041,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.142857142857142,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-331ec9a18e6c05e7daa0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ba6b9135c56c33d9b620946a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":533,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[266];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":533,"metadata":{"avgScore":0.40677645580102734,"avg_score":0.40677645580102734,"avg_score_percent":40.677645580102734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1626,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1796865182,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":6432918,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":143,"runtime":298162,"source_flags":[],"source_harness":"claude_code","source_row_index":266,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1626,"totalDurationS":298162,"totalInputTokens":1796865182,"totalOutputTokens":6432918,"total_cost_usd":1626,"total_input_tokens":1796865182,"total_output_tokens":6432918,"total_runtime_s":298162,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.36842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c7625b25e7cbe4b3cea3","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b34cef5f99f204b3a2532d88","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":637,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[318];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":637,"metadata":{"avgScore":0.18756758749923322,"avg_score":0.18756758749923322,"avg_score_percent":18.75675874992332,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":203.72,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":74904373,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1166124,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":103208,"source_flags":[],"source_harness":"claude_code","source_row_index":318,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":203.72,"totalDurationS":103208,"totalInputTokens":74904373,"totalOutputTokens":1166124,"total_cost_usd":203.72,"total_input_tokens":74904373,"total_output_tokens":1166124,"total_runtime_s":103208,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-56873b72fb4ce294434c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a635c865569adf94c2e76218","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2571428571428571,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":759,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[379];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":759,"metadata":{"avgScore":0.4406278923915453,"avg_score":0.4406278923915453,"avg_score_percent":44.06278923915453,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1953.22,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":642239869,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":15272399,"passRate":0.2571428571428571,"pass_rate":0.2571428571428571,"pass_rate_percent":25.71428571428571,"passes":27,"runs":104,"runtime":399408,"source_flags":[],"source_harness":"claude_code","source_row_index":379,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":1953.22,"totalDurationS":399408,"totalInputTokens":642239869,"totalOutputTokens":15272399,"total_cost_usd":1953.22,"total_input_tokens":642239869,"total_output_tokens":15272399,"total_runtime_s":399408,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.71428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-673e280b43bec306972e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a2d883cc50915377b6024213","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3582089552238806,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":403,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[201];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":403,"metadata":{"avgScore":0.6273164650789494,"avg_score":0.6273164650789494,"avg_score_percent":62.73164650789494,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":683.28,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":201294188,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5981020,"passRate":0.3582089552238806,"pass_rate":0.3582089552238806,"pass_rate_percent":35.82089552238806,"passes":24,"runs":66,"runtime":151680,"source_flags":[],"source_harness":"claude_code","source_row_index":201,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":683.28,"totalDurationS":151680,"totalInputTokens":201294188,"totalOutputTokens":5981020,"total_cost_usd":683.28,"total_input_tokens":201294188,"total_output_tokens":5981020,"total_runtime_s":151680,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.82089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a3ff2898c03676b0283d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a07c5918f3b25e7f2c5b8fa1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.373134328358209,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":427,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[213];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":427,"metadata":{"avgScore":0.598098576643749,"avg_score":0.598098576643749,"avg_score_percent":59.8098576643749,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":397673605,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":2241277,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":76504,"source_flags":[],"source_harness":"claude_code","source_row_index":213,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":446,"totalDurationS":76504,"totalInputTokens":397673605,"totalOutputTokens":2241277,"total_cost_usd":446,"total_input_tokens":397673605,"total_output_tokens":2241277,"total_runtime_s":76504,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.3134328358209,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-087a3691344d7657afec","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-93c3b778bef21c9b87d2d364","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.12,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1033,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[516];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1033,"metadata":{"avgScore":0.2984717724109289,"avg_score":0.2984717724109289,"avg_score_percent":29.847177241092893,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.82,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":97287341,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1049930,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":50,"runtime":159935,"source_flags":[],"source_harness":"claude_code","source_row_index":516,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":415.82,"totalDurationS":159935,"totalInputTokens":97287341,"totalOutputTokens":1049930,"total_cost_usd":415.82,"total_input_tokens":97287341,"total_output_tokens":1049930,"total_runtime_s":159935,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3899fb90e3700a58818d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8bcc47b27c004f5a84b7c0e3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1003,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[501];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1003,"metadata":{"avgScore":0.3793836589504913,"avg_score":0.3793836589504913,"avg_score_percent":37.93836589504913,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":777.77,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":255949496,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5279930,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":10,"runs":50,"runtime":115789,"source_flags":[],"source_harness":"claude_code","source_row_index":501,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":777.77,"totalDurationS":115789,"totalInputTokens":255949496,"totalOutputTokens":5279930,"total_cost_usd":777.77,"total_input_tokens":255949496,"total_output_tokens":5279930,"total_runtime_s":115789,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dfc99ba213f2900630c2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8694a9a682dce135b3348483","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1163,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[581];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1163,"metadata":{"avgScore":0.06638999999999999,"avg_score":0.06638999999999999,"avg_score_percent":6.638999999999999,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":158.25,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":32912598,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":542905,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":25,"runtime":174170,"source_flags":[],"source_harness":"claude_code","source_row_index":581,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":25,"tokens":null,"totalCostUsd":158.25,"totalDurationS":174170,"totalInputTokens":32912598,"totalOutputTokens":542905,"total_cost_usd":158.25,"total_input_tokens":32912598,"total_output_tokens":542905,"total_runtime_s":174170,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cb7f1a7d042da35e01f9","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8624d62b0cfa7472a187b0d6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.1773049645390071,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":909,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[454];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":909,"metadata":{"avgScore":0.4067685680771121,"avg_score":0.4067685680771121,"avg_score_percent":40.67685680771121,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1328.51,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":307300337,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":3412910,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":136,"runtime":466272,"source_flags":[],"source_harness":"claude_code","source_row_index":454,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1328.51,"totalDurationS":466272,"totalInputTokens":307300337,"totalOutputTokens":3412910,"total_cost_usd":1328.51,"total_input_tokens":307300337,"total_output_tokens":3412910,"total_runtime_s":466272,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.73049645390071,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b5a37982ed1bb33b569b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8176669573e383fb174e42f4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":617,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[308];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":617,"metadata":{"avgScore":0.22167032341406182,"avg_score":0.22167032341406182,"avg_score_percent":22.16703234140618,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":288.39,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":227246409,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1695241,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":101694,"source_flags":[],"source_harness":"claude_code","source_row_index":308,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":288.39,"totalDurationS":101694,"totalInputTokens":227246409,"totalOutputTokens":1695241,"total_cost_usd":288.39,"total_input_tokens":227246409,"total_output_tokens":1695241,"total_runtime_s":101694,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4d6e624ec014c580cbc5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7bf7ea797dab6707e524b66a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":329,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[164];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":329,"metadata":{"avgScore":0.052413157894736835,"avg_score":0.052413157894736835,"avg_score_percent":5.241315789473684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":601.73,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":157559357,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":870700,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":27,"runtime":186747,"source_flags":[],"source_harness":"claude_code","source_row_index":164,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":27,"tokens":null,"totalCostUsd":601.73,"totalDurationS":186747,"totalInputTokens":157559357,"totalOutputTokens":870700,"total_cost_usd":601.73,"total_input_tokens":157559357,"total_output_tokens":870700,"total_runtime_s":186747,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c1b47b5ef7349c691c0c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-788619d572d29f12a85a13be","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":61,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[30];split=full;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":61,"metadata":{"avgScore":0.40677645580102734,"avg_score":0.40677645580102734,"avg_score_percent":40.677645580102734,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1626,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":1796865182,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":6432918,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":143,"runtime":298162,"source_flags":[],"source_harness":"claude_code","source_row_index":30,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":1626,"totalDurationS":298162,"totalInputTokens":1796865182,"totalOutputTokens":6432918,"total_cost_usd":1626,"total_input_tokens":1796865182,"total_output_tokens":6432918,"total_runtime_s":298162,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.36842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-98ce461831c3677bd589","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-780db29a95f67c809c2ca3f1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":411,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[205];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":411,"metadata":{"avgScore":0.6172996937063634,"avg_score":0.6172996937063634,"avg_score_percent":61.72996937063634,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":805.69,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":190131175,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1966191,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":156896,"source_flags":[],"source_harness":"claude_code","source_row_index":205,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":805.69,"totalDurationS":156896,"totalInputTokens":190131175,"totalOutputTokens":1966191,"total_cost_usd":805.69,"total_input_tokens":190131175,"total_output_tokens":1966191,"total_runtime_s":156896,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-297a44570e03666e613c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-754c807b6f586bc791ecf4f4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.10909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":199,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[99];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":199,"metadata":{"avgScore":0.2873379749190263,"avg_score":0.2873379749190263,"avg_score_percent":28.733797491902628,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":485.88,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":118264612,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1117633,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":54,"runtime":171502,"source_flags":[],"source_harness":"claude_code","source_row_index":99,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":485.88,"totalDurationS":171502,"totalInputTokens":118264612,"totalOutputTokens":1117633,"total_cost_usd":485.88,"total_input_tokens":118264612,"total_output_tokens":1117633,"total_runtime_s":171502,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.909090909090908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a7439d22c862019062a0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-73358c3bdcf6aff9d210c6bf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":297,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[148];split=full/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":297,"metadata":{"avgScore":0.10431719138556748,"avg_score":0.10431719138556748,"avg_score_percent":10.431719138556748,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1395.12,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":739811872,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":5918841,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":38,"runtime":253733,"source_flags":[],"source_harness":"claude_code","source_row_index":148,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":1395.12,"totalDurationS":253733,"totalInputTokens":739811872,"totalOutputTokens":5918841,"total_cost_usd":1395.12,"total_input_tokens":739811872,"total_output_tokens":5918841,"total_runtime_s":253733,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4adfd4be0c79c234524b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7260fa261c30a35c0bddccdf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":993,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[496];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":993,"metadata":{"avgScore":0.3933536738539414,"avg_score":0.3933536738539414,"avg_score_percent":39.335367385394136,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":516.84,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":544268438,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1875055,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":10,"runs":49,"runtime":76526,"source_flags":[],"source_harness":"claude_code","source_row_index":496,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":516.84,"totalDurationS":76526,"totalInputTokens":544268438,"totalOutputTokens":1875055,"total_cost_usd":516.84,"total_input_tokens":544268438,"total_output_tokens":1875055,"total_runtime_s":76526,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4ae23d3dbed7f23b1d8f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-71865a5a89f6ddc342c6620f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1149,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[574];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1149,"metadata":{"avgScore":0.0856096444480286,"avg_score":0.0856096444480286,"avg_score_percent":8.56096444480286,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":580.14,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":726419627,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1662333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":26,"runtime":135036,"source_flags":[],"source_harness":"claude_code","source_row_index":574,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":580.14,"totalDurationS":135036,"totalInputTokens":726419627,"totalOutputTokens":1662333,"total_cost_usd":580.14,"total_input_tokens":726419627,"total_output_tokens":1662333,"total_runtime_s":135036,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c7802c472a8e5aad4a7c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-6d916245e5546af3aea6f1f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.1773049645390071,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1381,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[690];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1381,"metadata":{"avgScore":0.4067685680771121,"avg_score":0.4067685680771121,"avg_score_percent":40.67685680771121,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1328.51,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":307300337,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":3412910,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":136,"runtime":466272,"source_flags":[],"source_harness":"claude_code","source_row_index":690,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1328.51,"totalDurationS":466272,"totalInputTokens":307300337,"totalOutputTokens":3412910,"total_cost_usd":1328.51,"total_input_tokens":307300337,"total_output_tokens":3412910,"total_runtime_s":466272,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.73049645390071,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1a6b230c4882ef2b5ff4","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-6a4dbf8586278714e962f9b2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":509,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[254];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":509,"metadata":{"avgScore":0.4514857743894153,"avg_score":0.4514857743894153,"avg_score_percent":45.14857743894153,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3985,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1692408720,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":22067608,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":150,"runtime":645508,"source_flags":[],"source_harness":"claude_code","source_row_index":254,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":3985,"totalDurationS":645508,"totalInputTokens":1692408720,"totalOutputTokens":22067608,"total_cost_usd":3985,"total_input_tokens":1692408720,"total_output_tokens":22067608,"total_runtime_s":645508,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-52165ce42fba1adbdd2d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-6074c8d01555748b3fd9817f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":521,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[260];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":521,"metadata":{"avgScore":0.4234878390395742,"avg_score":0.4234878390395742,"avg_score_percent":42.34878390395742,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2815,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1209502455,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":16507476,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":151,"runtime":519320,"source_flags":[],"source_harness":"claude_code","source_row_index":260,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":2815,"totalDurationS":519320,"totalInputTokens":1209502455,"totalOutputTokens":16507476,"total_cost_usd":2815,"total_input_tokens":1209502455,"total_output_tokens":16507476,"total_runtime_s":519320,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.36842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e441a88258903c7a14ea","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5b67a79bd65009e5d0b27814","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":887,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[443];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":887,"metadata":{"avgScore":0.44189296504626746,"avg_score":0.44189296504626746,"avg_score_percent":44.189296504626746,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2611.28,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1134598082,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":15341352,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":140,"runtime":416112,"source_flags":[],"source_harness":"claude_code","source_row_index":443,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":2611.28,"totalDurationS":416112,"totalInputTokens":1134598082,"totalOutputTokens":15341352,"total_cost_usd":2611.28,"total_input_tokens":1134598082,"total_output_tokens":15341352,"total_runtime_s":416112,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-01ea06b5f7a824092ef9","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5136adf3983edf30ae25c6ab","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.43283582089552236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1237,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[618];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1237,"metadata":{"avgScore":0.6396884267777608,"avg_score":0.6396884267777608,"avg_score_percent":63.96884267777608,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1105.97,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":310513347,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":8274090,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":66,"runtime":178788,"source_flags":[],"source_harness":"claude_code","source_row_index":618,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":1105.97,"totalDurationS":178788,"totalInputTokens":310513347,"totalOutputTokens":8274090,"total_cost_usd":1105.97,"total_input_tokens":310513347,"total_output_tokens":8274090,"total_runtime_s":178788,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.28358208955223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-119a2906986d56b43f53","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-460ec2ecd65f7a97bcebe3fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.16447368421052633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":547,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[273];split=full/overall;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":547,"metadata":{"avgScore":0.38312084275574215,"avg_score":0.38312084275574215,"avg_score_percent":38.312084275574215,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1837.79,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":452042179,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":3793251,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":141,"runtime":489984,"source_flags":[],"source_harness":"claude_code","source_row_index":273,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":1837.79,"totalDurationS":489984,"totalInputTokens":452042179,"totalOutputTokens":3793251,"total_cost_usd":1837.79,"total_input_tokens":452042179,"total_output_tokens":3793251,"total_runtime_s":489984,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.447368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9a1e55328a5893e53c89","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-42a40fa455d4115a4e9e54b2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":671,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[335];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":671,"metadata":{"avgScore":0.08,"avg_score":0.08,"avg_score_percent":8.0,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":509.28,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":144741842,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":380341,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":23712,"source_flags":[],"source_harness":"claude_code","source_row_index":335,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":509.28,"totalDurationS":23712,"totalInputTokens":144741842,"totalOutputTokens":380341,"total_cost_usd":509.28,"total_input_tokens":144741842,"total_output_tokens":380341,"total_runtime_s":23712,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0b699693120813c71854","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-36d5da8d39be86be85350a72","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":137,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[68];split=full/full-spectrum;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":137,"metadata":{"avgScore":0.4313229586002382,"avg_score":0.4313229586002382,"avg_score_percent":43.132295860023824,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1323.79,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":524284122,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":8213812,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":55,"runtime":214602,"source_flags":[],"source_harness":"claude_code","source_row_index":68,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":1323.79,"totalDurationS":214602,"totalInputTokens":524284122,"totalOutputTokens":8213812,"total_cost_usd":1323.79,"total_input_tokens":524284122,"total_output_tokens":8213812,"total_runtime_s":214602,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.636363636363637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-46085d14a078ed3b65b9","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2e5fbb9e93ffd3357e982912","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":775,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[387];split=linux_only;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":775,"metadata":{"avgScore":0.42142958861655383,"avg_score":0.42142958861655383,"avg_score_percent":42.14295886165538,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1713.22,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":640997605,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":11872346,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":21,"runs":104,"runtime":304614,"source_flags":[],"source_harness":"claude_code","source_row_index":387,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":1713.22,"totalDurationS":304614,"totalInputTokens":640997605,"totalOutputTokens":11872346,"total_cost_usd":1713.22,"total_input_tokens":640997605,"total_output_tokens":11872346,"total_runtime_s":304614,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c627b939fbdb75bb76b3","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2b246f930d605522eab34d07","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-high","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":663,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[331];split=licensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":663,"metadata":{"avgScore":0.1083130564079132,"avg_score":0.1083130564079132,"avg_score_percent":10.831305640791319,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":117.67,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-high","input_tokens":225487387,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":811747,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":8,"runtime":30318,"source_flags":[],"source_harness":"claude_code","source_row_index":331,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":8,"tokens":null,"totalCostUsd":117.67,"totalDurationS":30318,"totalInputTokens":225487387,"totalOutputTokens":811747,"total_cost_usd":117.67,"total_input_tokens":225487387,"total_output_tokens":811747,"total_runtime_s":30318,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0c0254c19af79f7c1932","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2ad970c925746cbc67a0af9e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1359,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[679];split=unlicensed/overall;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1359,"metadata":{"avgScore":0.44189296504626746,"avg_score":0.44189296504626746,"avg_score_percent":44.189296504626746,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2611.28,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1134598082,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":15341352,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":140,"runtime":416112,"source_flags":[],"source_harness":"claude_code","source_row_index":679,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":2611.28,"totalDurationS":416112,"totalInputTokens":1134598082,"totalOutputTokens":15341352,"total_cost_usd":2611.28,"total_input_tokens":1134598082,"total_output_tokens":15341352,"total_runtime_s":416112,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-96b15d2d8e113a431ce5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2611c5c9f703e2b727460b2c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":37,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[18];split=full;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":37,"metadata":{"avgScore":0.4514857743894153,"avg_score":0.4514857743894153,"avg_score_percent":45.14857743894153,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3985,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1692408720,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":22067608,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":150,"runtime":645508,"source_flags":[],"source_harness":"claude_code","source_row_index":18,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":150,"tokens":null,"totalCostUsd":3985,"totalDurationS":645508,"totalInputTokens":1692408720,"totalOutputTokens":22067608,"total_cost_usd":3985,"total_input_tokens":1692408720,"total_output_tokens":22067608,"total_runtime_s":645508,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-509b84e0d63eebb33f94","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-23a3cb2eeeea3a2c0f2f4832","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":49,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[24];split=full;harness=claude_code;model=claude-opus-4-8;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":49,"metadata":{"avgScore":0.4234878390395742,"avg_score":0.4234878390395742,"avg_score_percent":42.34878390395742,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2815,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1209502455,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":16507476,"passRate":0.2236842105263158,"pass_rate":0.2236842105263158,"pass_rate_percent":22.36842105263158,"passes":34,"runs":151,"runtime":519320,"source_flags":[],"source_harness":"claude_code","source_row_index":24,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":2815,"totalDurationS":519320,"totalInputTokens":1209502455,"totalOutputTokens":16507476,"total_cost_usd":2815,"total_input_tokens":1209502455,"total_output_tokens":16507476,"total_runtime_s":519320,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.36842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7954f95ba9cfa777e83b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-23385641e7db5367e04a9161","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official+adj","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":null,"source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1247,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[623];split=unlicensed/near-term;harness=claude_code;model=claude-opus-4-8;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1247,"metadata":{"avgScore":0.6172996937063634,"avg_score":0.6172996937063634,"avg_score_percent":61.72996937063634,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":805.69,"costSource":"official+adj","cost_source":"official+adj","harness":"claude_code","harness_id":"claude-code","harness_variant":null,"input_tokens":190131175,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":1966191,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":67,"runtime":156896,"source_flags":[],"source_harness":"claude_code","source_row_index":623,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":805.69,"totalDurationS":156896,"totalInputTokens":190131175,"totalOutputTokens":1966191,"total_cost_usd":805.69,"total_input_tokens":190131175,"total_output_tokens":1966191,"total_runtime_s":156896,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f4919cc9accb624a41c8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0d90c50b6e602ad5569feb7e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.43283582089552236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":401,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[200];split=full/near-term;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":401,"metadata":{"avgScore":0.6396884267777608,"avg_score":0.6396884267777608,"avg_score_percent":63.96884267777608,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1105.97,"costSource":"published","cost_source":"published","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":310513347,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":8274090,"passRate":0.43283582089552236,"pass_rate":0.43283582089552236,"pass_rate_percent":43.28358208955223,"passes":29,"runs":66,"runtime":178788,"source_flags":[],"source_harness":"claude_code","source_row_index":200,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":1105.97,"totalDurationS":178788,"totalInputTokens":310513347,"totalOutputTokens":8274090,"total_cost_usd":1105.97,"total_input_tokens":310513347,"total_output_tokens":8274090,"total_runtime_s":178788,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.28358208955223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":48,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4b9e9de4e27efac61bc2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0a4a8af126375116f7287329","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"published+task-ratio","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2907801418439716,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"614a48db911e736898422420016b95602b937ecf","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":873,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[436];split=unlicensed;harness=claude_code;model=claude-opus-4-8;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":873,"metadata":{"avgScore":0.4694146393591237,"avg_score":0.4694146393591237,"avg_score_percent":46.94146393591237,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3696.61,"costSource":"published+task-ratio","cost_source":"published+task-ratio","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1465162311,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-4-8","output_tokens":20372367,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":139,"runtime":543814,"source_flags":[],"source_harness":"claude_code","source_row_index":436,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":3696.61,"totalDurationS":543814,"totalInputTokens":1465162311,"totalOutputTokens":20372367,"total_cost_usd":3696.61,"total_input_tokens":1465162311,"total_output_tokens":20372367,"total_runtime_s":543814,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.078014184397162,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e43a5b04618ac06fcff1","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-fc5a040bda11cbda66ca16b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.49001017760022275,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":496,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[247];split=full/overall;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":496,"metadata":{"avgScore":0.49001017760022275,"avg_score":0.49001017760022275,"avg_score_percent":49.00101776002227,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":7239.27,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1802968693,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":12892839,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":636862.3549999996,"source_flags":[],"source_harness":"claude_code","source_row_index":247,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":7239.27,"totalDurationS":636862.3549999996,"totalInputTokens":1802968693,"totalOutputTokens":12892839,"total_cost_usd":7239.27,"total_input_tokens":1802968693,"total_output_tokens":12892839,"total_runtime_s":636862.3549999996,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.00101776002227,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fa58dcd6ada2d4514630","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e2d625ac29104675d8f1510f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4854044140124046,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":736,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[367];split=linux_only;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":736,"metadata":{"avgScore":0.4854044140124046,"avg_score":0.4854044140124046,"avg_score_percent":48.54044140124046,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3808.99,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":738738178,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":8290040,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":105,"runtime":340170.9999999999,"source_flags":[],"source_harness":"claude_code","source_row_index":367,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":3808.99,"totalDurationS":340170.9999999999,"totalInputTokens":738738178,"totalOutputTokens":8290040,"total_cost_usd":3808.99,"total_input_tokens":738738178,"total_output_tokens":8290040,"total_runtime_s":340170.9999999999,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.54044140124046,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6bcf7ae95bf13e769e26","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-cb599a9aad744f856f417424","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.504046071947629,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":862,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[430];split=unlicensed;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":862,"metadata":{"avgScore":0.504046071947629,"avg_score":0.504046071947629,"avg_score_percent":50.404607194762896,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":6348.02,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1342300803,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":11550039,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":141,"runtime":540095.1779999996,"source_flags":[],"source_harness":"claude_code","source_row_index":430,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":6348.02,"totalDurationS":540095.1779999996,"totalInputTokens":1342300803,"totalOutputTokens":11550039,"total_cost_usd":6348.02,"total_input_tokens":1342300803,"total_output_tokens":11550039,"total_runtime_s":540095.1779999996,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.404607194762896,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0118999a42a3f24ed6ee","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b8125d8e1d992a9362d04163","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.39488395328769904,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":990,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[494];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":990,"metadata":{"avgScore":0.39488395328769904,"avg_score":0.39488395328769904,"avg_score_percent":39.488395328769904,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1927.35,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":377325112,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":3910682,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":50,"runtime":148755.5,"source_flags":[],"source_harness":"claude_code","source_row_index":494,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":1927.35,"totalDurationS":148755.5,"totalInputTokens":377325112,"totalOutputTokens":3910682,"total_cost_usd":1927.35,"total_input_tokens":377325112,"total_output_tokens":3910682,"total_runtime_s":148755.5,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.488395328769904,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-66cc1f93192c5e8b00a3","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-91fc912ecc129c75467e0255","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.504046071947629,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1334,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[666];split=unlicensed/overall;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1334,"metadata":{"avgScore":0.504046071947629,"avg_score":0.504046071947629,"avg_score_percent":50.404607194762896,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":6348.02,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1342300803,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":11550039,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":141,"runtime":540095.1779999996,"source_flags":[],"source_harness":"claude_code","source_row_index":666,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":6348.02,"totalDurationS":540095.1779999996,"totalInputTokens":1342300803,"totalOutputTokens":11550039,"total_cost_usd":6348.02,"total_input_tokens":1342300803,"total_output_tokens":11550039,"total_runtime_s":540095.1779999996,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.404607194762896,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-34cbfad76b3f02be4dfd","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4ed8a0557259c78ed0541c2f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6806900632972314,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1222,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[610];split=unlicensed/near-term;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1222,"metadata":{"avgScore":0.6806900632972314,"avg_score":0.6806900632972314,"avg_score_percent":68.06900632972314,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1977.73,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":517910640,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":4443900,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":175785.86699999997,"source_flags":[],"source_harness":"claude_code","source_row_index":610,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1977.73,"totalDurationS":175785.86699999997,"totalInputTokens":517910640,"totalOutputTokens":4443900,"total_cost_usd":1977.73,"total_input_tokens":517910640,"total_output_tokens":4443900,"total_runtime_s":175785.86699999997,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":68.06900632972314,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0fe8ee8c8ada399f19a0","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3f96bcde836197240d2f9888","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.31009553187437733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":596,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[297];split=licensed;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":596,"metadata":{"avgScore":0.31009553187437733,"avg_score":0.31009553187437733,"avg_score_percent":31.00955318743773,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":891.25,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":460667890,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":1342800,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":10,"runtime":96767.177,"source_flags":[],"source_harness":"claude_code","source_row_index":297,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":891.25,"totalDurationS":96767.177,"totalInputTokens":460667890,"totalOutputTokens":1342800,"total_cost_usd":891.25,"total_input_tokens":460667890,"total_output_tokens":1342800,"total_runtime_s":96767.177,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.00955318743773,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f64ab1e8cdc6781d1e31","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3c95019adb2a0aee3a709b2f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6806900632972314,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":386,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[192];split=full/near-term;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":386,"metadata":{"avgScore":0.6806900632972314,"avg_score":0.6806900632972314,"avg_score_percent":68.06900632972314,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1977.73,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":517910640,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":4443900,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":175785.86699999997,"source_flags":[],"source_harness":"claude_code","source_row_index":192,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1977.73,"totalDurationS":175785.86699999997,"totalInputTokens":517910640,"totalOutputTokens":4443900,"total_cost_usd":1977.73,"total_input_tokens":517910640,"total_output_tokens":4443900,"total_runtime_s":175785.86699999997,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":68.06900632972314,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e407e6503e0e32d4f2f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-31fa40442a1ba39c2461af61","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.39462177571609003,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":156,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[77];split=full/full-spectrum;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":156,"metadata":{"avgScore":0.39462177571609003,"avg_score":0.39462177571609003,"avg_score_percent":39.462177571609004,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2249.64,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":596096745,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":4409259,"passRate":0.21818181818181817,"pass_rate":0.21818181818181817,"pass_rate_percent":21.818181818181817,"passes":12,"runs":55,"runtime":186927.14900000003,"source_flags":[],"source_harness":"claude_code","source_row_index":77,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":2249.64,"totalDurationS":186927.14900000003,"totalInputTokens":596096745,"totalOutputTokens":4409259,"total_cost_usd":2249.64,"total_input_tokens":596096745,"total_output_tokens":4409259,"total_runtime_s":186927.14900000003,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.462177571609004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-347e9f2b027c604b7069","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2d585483eaa995a5de4701b8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23266092341932604,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":242,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[120];split=full/last-exam;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":242,"metadata":{"avgScore":0.23266092341932604,"avg_score":0.23266092341932604,"avg_score_percent":23.266092341932605,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3850.8,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":905477605,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":5061105,"passRate":0.07894736842105263,"pass_rate":0.07894736842105263,"pass_rate_percent":7.894736842105263,"passes":3,"runs":37,"runtime":331459.134,"source_flags":[],"source_harness":"claude_code","source_row_index":120,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":3850.8,"totalDurationS":331459.134,"totalInputTokens":905477605,"totalOutputTokens":5061105,"total_cost_usd":3850.8,"total_input_tokens":905477605,"total_output_tokens":5061105,"total_runtime_s":331459.134,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.266092341932605,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-65fd35232deca1f5cc4a","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1c52caed82e6a9c3989a9592","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24633547464387465,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1076,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[537];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1076,"metadata":{"avgScore":0.24633547464387465,"avg_score":0.24633547464387465,"avg_score_percent":24.633547464387465,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3253.61,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":579520081,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":4133743,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":30,"runtime":251013.41100000002,"source_flags":[],"source_harness":"claude_code","source_row_index":537,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":3253.61,"totalDurationS":251013.41100000002,"totalInputTokens":579520081,"totalOutputTokens":4133743,"total_cost_usd":3253.61,"total_input_tokens":579520081,"total_output_tokens":4133743,"total_runtime_s":251013.41100000002,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.633547464387465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d837fff74cd5c41f0ba","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1b19721de4ffe82a84ede87a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.49001017760022275,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f16d733349b5f0c754f58b9c7bd60beddef1579","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":24,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[11];split=full;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":24,"metadata":{"avgScore":0.49001017760022275,"avg_score":0.49001017760022275,"avg_score_percent":49.00101776002227,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":7239.27,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1802968693,"metric_id":"avg_score","metric_role":"secondary","model":"claude-opus-5","output_tokens":12892839,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":636862.3549999996,"source_flags":[],"source_harness":"claude_code","source_row_index":11,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":7239.27,"totalDurationS":636862.3549999996,"totalInputTokens":1802968693,"totalOutputTokens":12892839,"total_cost_usd":7239.27,"total_input_tokens":1802968693,"total_output_tokens":12892839,"total_runtime_s":636862.3549999996,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.00101776002227,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd2b6bb41e2410c16c46","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d9354f8c477430b2f5a9eb52","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28368794326241137,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1333,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[666];split=unlicensed/overall;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1333,"metadata":{"avgScore":0.504046071947629,"avg_score":0.504046071947629,"avg_score_percent":50.404607194762896,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":6348.02,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1342300803,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":11550039,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":141,"runtime":540095.1779999996,"source_flags":[],"source_harness":"claude_code","source_row_index":666,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":6348.02,"totalDurationS":540095.1779999996,"totalInputTokens":1342300803,"totalOutputTokens":11550039,"total_cost_usd":6348.02,"total_input_tokens":1342300803,"total_output_tokens":11550039,"total_runtime_s":540095.1779999996,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.368794326241137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6cdffc0d571d47b1158b","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b056d40d170aa4dd7c2fa4f6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1075,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[537];split=unlicensed/last-exam;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1075,"metadata":{"avgScore":0.24633547464387465,"avg_score":0.24633547464387465,"avg_score_percent":24.633547464387465,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3253.61,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":579520081,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":4133743,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":30,"runtime":251013.41100000002,"source_flags":[],"source_harness":"claude_code","source_row_index":537,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":3253.61,"totalDurationS":251013.41100000002,"totalInputTokens":579520081,"totalOutputTokens":4133743,"total_cost_usd":3253.61,"total_input_tokens":579520081,"total_output_tokens":4133743,"total_runtime_s":251013.41100000002,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8b7db579162367ff3c0","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-9c320dc2c280f605ac447efc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28368794326241137,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":861,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[430];split=unlicensed;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":861,"metadata":{"avgScore":0.504046071947629,"avg_score":0.504046071947629,"avg_score_percent":50.404607194762896,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":6348.02,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1342300803,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":11550039,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":141,"runtime":540095.1779999996,"source_flags":[],"source_harness":"claude_code","source_row_index":430,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":6348.02,"totalDurationS":540095.1779999996,"totalInputTokens":1342300803,"totalOutputTokens":11550039,"total_cost_usd":6348.02,"total_input_tokens":1342300803,"total_output_tokens":11550039,"total_runtime_s":540095.1779999996,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.368794326241137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b2cad04cf843c2b9fc86","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8d2ba5f642c5059b648febce","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.21818181818181817,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":155,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[77];split=full/full-spectrum;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":155,"metadata":{"avgScore":0.39462177571609003,"avg_score":0.39462177571609003,"avg_score_percent":39.462177571609004,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2249.64,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":596096745,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":4409259,"passRate":0.21818181818181817,"pass_rate":0.21818181818181817,"pass_rate_percent":21.818181818181817,"passes":12,"runs":55,"runtime":186927.14900000003,"source_flags":[],"source_harness":"claude_code","source_row_index":77,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":2249.64,"totalDurationS":186927.14900000003,"totalInputTokens":596096745,"totalOutputTokens":4409259,"total_cost_usd":2249.64,"total_input_tokens":596096745,"total_output_tokens":4409259,"total_runtime_s":186927.14900000003,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.818181818181817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-910e2a1cded99655b7cf","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-8cc7383dc9a67a9c795d3991","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":495,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[247];split=full/overall;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":495,"metadata":{"avgScore":0.49001017760022275,"avg_score":0.49001017760022275,"avg_score_percent":49.00101776002227,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":7239.27,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1802968693,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":12892839,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":636862.3549999996,"source_flags":[],"source_harness":"claude_code","source_row_index":247,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":7239.27,"totalDurationS":636862.3549999996,"totalInputTokens":1802968693,"totalOutputTokens":12892839,"total_cost_usd":7239.27,"total_input_tokens":1802968693,"total_output_tokens":12892839,"total_runtime_s":636862.3549999996,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-60b11e11c40107333224","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-86cfebca257f76251ae8406a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40298507462686567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1221,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[610];split=unlicensed/near-term;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1221,"metadata":{"avgScore":0.6806900632972314,"avg_score":0.6806900632972314,"avg_score_percent":68.06900632972314,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1977.73,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":517910640,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":4443900,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":175785.86699999997,"source_flags":[],"source_harness":"claude_code","source_row_index":610,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1977.73,"totalDurationS":175785.86699999997,"totalInputTokens":517910640,"totalOutputTokens":4443900,"total_cost_usd":1977.73,"total_input_tokens":517910640,"total_output_tokens":4443900,"total_runtime_s":175785.86699999997,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.298507462686565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9e9c0c29912040294686","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-642643582fa557bc02543966","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":595,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[297];split=licensed;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":595,"metadata":{"avgScore":0.31009553187437733,"avg_score":0.31009553187437733,"avg_score_percent":31.00955318743773,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":891.25,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":460667890,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":1342800,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":10,"runtime":96767.177,"source_flags":[],"source_harness":"claude_code","source_row_index":297,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":891.25,"totalDurationS":96767.177,"totalInputTokens":460667890,"totalOutputTokens":1342800,"total_cost_usd":891.25,"total_input_tokens":460667890,"total_output_tokens":1342800,"total_runtime_s":96767.177,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a94c6c46efad0f52d704","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-561d1ed4e4040f77fbb77b08","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.24761904761904763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":735,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[367];split=linux_only;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":735,"metadata":{"avgScore":0.4854044140124046,"avg_score":0.4854044140124046,"avg_score_percent":48.54044140124046,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3808.99,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":738738178,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":8290040,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":105,"runtime":340170.9999999999,"source_flags":[],"source_harness":"claude_code","source_row_index":367,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":3808.99,"totalDurationS":340170.9999999999,"totalInputTokens":738738178,"totalOutputTokens":8290040,"total_cost_usd":3808.99,"total_input_tokens":738738178,"total_output_tokens":8290040,"total_runtime_s":340170.9999999999,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.761904761904763,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f4597f9a4029fa681734","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4cfd98e1dabf387f359d8857","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.07894736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":241,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[120];split=full/last-exam;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":241,"metadata":{"avgScore":0.23266092341932604,"avg_score":0.23266092341932604,"avg_score_percent":23.266092341932605,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":3850.8,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":905477605,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":5061105,"passRate":0.07894736842105263,"pass_rate":0.07894736842105263,"pass_rate_percent":7.894736842105263,"passes":3,"runs":37,"runtime":331459.134,"source_flags":[],"source_harness":"claude_code","source_row_index":120,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":3850.8,"totalDurationS":331459.134,"totalInputTokens":905477605,"totalOutputTokens":5061105,"total_cost_usd":3850.8,"total_input_tokens":905477605,"total_output_tokens":5061105,"total_runtime_s":331459.134,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.894736842105263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d2a404c46bf2fef3ad22","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1c73ba5764c7b6b1b0624102","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":23,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[11];split=full;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":23,"metadata":{"avgScore":0.49001017760022275,"avg_score":0.49001017760022275,"avg_score_percent":49.00101776002227,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":7239.27,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":1802968693,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":12892839,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":636862.3549999996,"source_flags":[],"source_harness":"claude_code","source_row_index":11,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":7239.27,"totalDurationS":636862.3549999996,"totalInputTokens":1802968693,"totalOutputTokens":12892839,"total_cost_usd":7239.27,"total_input_tokens":1802968693,"total_output_tokens":12892839,"total_runtime_s":636862.3549999996,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-01bc89941fcb6404bf58","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-15da7fa7e9e32a9ef5298ddd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":989,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[494];split=unlicensed/full-spectrum;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":989,"metadata":{"avgScore":0.39488395328769904,"avg_score":0.39488395328769904,"avg_score_percent":39.488395328769904,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1927.35,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":377325112,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":3910682,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":50,"runtime":148755.5,"source_flags":[],"source_harness":"claude_code","source_row_index":494,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":1927.35,"totalDurationS":148755.5,"totalInputTokens":377325112,"totalOutputTokens":3910682,"total_cost_usd":1927.35,"total_input_tokens":377325112,"total_output_tokens":3910682,"total_runtime_s":148755.5,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-79ce1bab847eb87e7925","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0970201dd734938637ca3bfb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40298507462686567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44d0f9700af4ee873682c419149a9b2dd6b07b9f","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":385,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[192];split=full/near-term;harness=claude_code;model=claude-opus-5;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":385,"metadata":{"avgScore":0.6806900632972314,"avg_score":0.6806900632972314,"avg_score_percent":68.06900632972314,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1977.73,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":517910640,"metric_id":"pass_rate","metric_role":"primary","model":"claude-opus-5","output_tokens":4443900,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":175785.86699999997,"source_flags":[],"source_harness":"claude_code","source_row_index":192,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":1977.73,"totalDurationS":175785.86699999997,"totalInputTokens":517910640,"totalOutputTokens":4443900,"total_cost_usd":1977.73,"total_input_tokens":517910640,"total_output_tokens":4443900,"total_runtime_s":175785.86699999997,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.298507462686565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8a3506755a21f2669ca8","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-abe5aec50a114ebf4afdfe38","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"forgecode","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.28997250926121826,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f89a03dc7b2e287fd8114eac9ab42f8eabd41721","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":818,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[408];split=linux_only;harness=forgecode;model=claude-sonnet-4-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":818,"metadata":{"avgScore":0.28997250926121826,"avg_score":0.28997250926121826,"avg_score_percent":28.997250926121826,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":100.05,"costSource":"official","cost_source":"official","harness":"forgecode","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":136530209.83333334,"metric_id":"avg_score","metric_role":"secondary","model":"claude-sonnet-4-6","output_tokens":1726369,"passRate":0.13333333333333336,"pass_rate":0.13333333333333336,"pass_rate_percent":13.333333333333336,"passes":38,"runs":244,"runtime":185485.77300154004,"source_flags":[],"source_harness":"forgecode","source_row_index":408,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":97,"tokens":null,"totalCostUsd":100.05,"totalDurationS":185485.77300154004,"totalInputTokens":136530209.83333334,"totalOutputTokens":1726369,"total_cost_usd":100.05,"total_input_tokens":136530209.83333334,"total_output_tokens":1726369,"total_runtime_s":185485.77300154004,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.997250926121826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6eca7fea3bd053d0f643","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6f72b75326fb1bd44ede427e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"hermes","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3203670919271656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f89a03dc7b2e287fd8114eac9ab42f8eabd41721","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":808,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[403];split=linux_only;harness=hermes;model=claude-sonnet-4-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":808,"metadata":{"avgScore":0.3203670919271656,"avg_score":0.3203670919271656,"avg_score_percent":32.036709192716565,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":436.6,"costSource":"official","cost_source":"official","harness":"hermes","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":195596773.66666666,"metric_id":"avg_score","metric_role":"secondary","model":"claude-sonnet-4-6","output_tokens":2001710.3333333333,"passRate":0.13174603174603172,"pass_rate":0.13174603174603172,"pass_rate_percent":13.174603174603172,"passes":36,"runs":248,"runtime":90692.90777800296,"source_flags":[],"source_harness":"hermes","source_row_index":403,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":96,"tokens":null,"totalCostUsd":436.6,"totalDurationS":90692.90777800296,"totalInputTokens":195596773.66666666,"totalOutputTokens":2001710.3333333333,"total_cost_usd":436.6,"total_input_tokens":195596773.66666666,"total_output_tokens":2001710.3333333333,"total_runtime_s":90692.90777800296,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.036709192716565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-97877b98438b924929f7","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6149775eb638501571fa335b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3164026085498648,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f89a03dc7b2e287fd8114eac9ab42f8eabd41721","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":810,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[404];split=linux_only;harness=openclaw;model=claude-sonnet-4-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":810,"metadata":{"avgScore":0.3164026085498648,"avg_score":0.3164026085498648,"avg_score_percent":31.64026085498648,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":181.33,"costSource":"official","cost_source":"official","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":247046472,"metric_id":"avg_score","metric_role":"secondary","model":"claude-sonnet-4-6","output_tokens":1875276,"passRate":0.11428571428571428,"pass_rate":0.11428571428571428,"pass_rate_percent":11.428571428571429,"passes":12,"runs":88,"runtime":121852.04607221007,"source_flags":[],"source_harness":"openclaw","source_row_index":404,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":88,"tokens":null,"totalCostUsd":181.33,"totalDurationS":121852.04607221007,"totalInputTokens":247046472,"totalOutputTokens":1875276,"total_cost_usd":181.33,"total_input_tokens":247046472,"total_output_tokens":1875276,"total_runtime_s":121852.04607221007,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.64026085498648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5d7833940392477b33dd","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2a32534c04705c623ac670a6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"terminus_2","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.30942664987397717,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f89a03dc7b2e287fd8114eac9ab42f8eabd41721","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":812,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[405];split=linux_only;harness=terminus_2;model=claude-sonnet-4-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":812,"metadata":{"avgScore":0.30942664987397717,"avg_score":0.30942664987397717,"avg_score_percent":30.942664987397716,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":319.72,"costSource":"official","cost_source":"official","harness":"terminus_2","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":760392746.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-sonnet-4-6","output_tokens":3129737,"passRate":0.11904761904761904,"pass_rate":0.11904761904761904,"pass_rate_percent":11.904761904761903,"passes":24,"runs":179,"runtime":262860.15114356764,"source_flags":[],"source_harness":"terminus_2","source_row_index":405,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":97,"tokens":null,"totalCostUsd":319.72,"totalDurationS":262860.15114356764,"totalInputTokens":760392746.5,"totalOutputTokens":3129737,"total_cost_usd":319.72,"total_input_tokens":760392746.5,"total_output_tokens":3129737,"total_runtime_s":262860.15114356764,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":30.942664987397716,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-939264235661c9bf9716","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0ce78066f882ada7830faadb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openhands","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19774111108214873,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f89a03dc7b2e287fd8114eac9ab42f8eabd41721","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":828,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[413];split=linux_only;harness=openhands;model=claude-sonnet-4-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":828,"metadata":{"avgScore":0.19774111108214873,"avg_score":0.19774111108214873,"avg_score_percent":19.77411110821487,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":242.67,"costSource":"official","cost_source":"official","harness":"openhands","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":349872550.5,"metric_id":"avg_score","metric_role":"secondary","model":"claude-sonnet-4-6","output_tokens":4376178.5,"passRate":0.09047619047619047,"pass_rate":0.09047619047619047,"pass_rate_percent":9.047619047619047,"passes":28,"runs":285,"runtime":128828.19181078624,"source_flags":[],"source_harness":"openhands","source_row_index":413,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":97,"tokens":null,"totalCostUsd":242.67,"totalDurationS":128828.19181078624,"totalInputTokens":349872550.5,"totalOutputTokens":4376178.5,"total_cost_usd":242.67,"total_input_tokens":349872550.5,"total_output_tokens":4376178.5,"total_runtime_s":128828.19181078624,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.77411110821487,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a2ea3e01732bbd0b3df7","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dc71988a2c87829cdf2134fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"terminus_2","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11904761904761904,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29ca005bffc2b99ad8d0e101b996221b8ecc2df8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":811,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[405];split=linux_only;harness=terminus_2;model=claude-sonnet-4-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":811,"metadata":{"avgScore":0.30942664987397717,"avg_score":0.30942664987397717,"avg_score_percent":30.942664987397716,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":319.72,"costSource":"official","cost_source":"official","harness":"terminus_2","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":760392746.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-sonnet-4-6","output_tokens":3129737,"passRate":0.11904761904761904,"pass_rate":0.11904761904761904,"pass_rate_percent":11.904761904761903,"passes":24,"runs":179,"runtime":262860.15114356764,"source_flags":[],"source_harness":"terminus_2","source_row_index":405,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":97,"tokens":null,"totalCostUsd":319.72,"totalDurationS":262860.15114356764,"totalInputTokens":760392746.5,"totalOutputTokens":3129737,"total_cost_usd":319.72,"total_input_tokens":760392746.5,"total_output_tokens":3129737,"total_runtime_s":262860.15114356764,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.904761904761903,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a47919338568c561b743","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7607861a779bedfb5ffb6845","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11428571428571428,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29ca005bffc2b99ad8d0e101b996221b8ecc2df8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":809,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[404];split=linux_only;harness=openclaw;model=claude-sonnet-4-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":809,"metadata":{"avgScore":0.3164026085498648,"avg_score":0.3164026085498648,"avg_score_percent":31.64026085498648,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":181.33,"costSource":"official","cost_source":"official","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":247046472,"metric_id":"pass_rate","metric_role":"primary","model":"claude-sonnet-4-6","output_tokens":1875276,"passRate":0.11428571428571428,"pass_rate":0.11428571428571428,"pass_rate_percent":11.428571428571429,"passes":12,"runs":88,"runtime":121852.04607221007,"source_flags":[],"source_harness":"openclaw","source_row_index":404,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":88,"tokens":null,"totalCostUsd":181.33,"totalDurationS":121852.04607221007,"totalInputTokens":247046472,"totalOutputTokens":1875276,"total_cost_usd":181.33,"total_input_tokens":247046472,"total_output_tokens":1875276,"total_runtime_s":121852.04607221007,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.428571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4e34fd686caca51fea15","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-560024e5a6a2dc290450d9de","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"forgecode","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13333333333333336,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29ca005bffc2b99ad8d0e101b996221b8ecc2df8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":817,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[408];split=linux_only;harness=forgecode;model=claude-sonnet-4-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":817,"metadata":{"avgScore":0.28997250926121826,"avg_score":0.28997250926121826,"avg_score_percent":28.997250926121826,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":100.05,"costSource":"official","cost_source":"official","harness":"forgecode","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":136530209.83333334,"metric_id":"pass_rate","metric_role":"primary","model":"claude-sonnet-4-6","output_tokens":1726369,"passRate":0.13333333333333336,"pass_rate":0.13333333333333336,"pass_rate_percent":13.333333333333336,"passes":38,"runs":244,"runtime":185485.77300154004,"source_flags":[],"source_harness":"forgecode","source_row_index":408,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":97,"tokens":null,"totalCostUsd":100.05,"totalDurationS":185485.77300154004,"totalInputTokens":136530209.83333334,"totalOutputTokens":1726369,"total_cost_usd":100.05,"total_input_tokens":136530209.83333334,"total_output_tokens":1726369,"total_runtime_s":185485.77300154004,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.333333333333336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8980c9b25c713c08ed4","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-32e756bdc29d8d4fedb30150","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openhands","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09047619047619047,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29ca005bffc2b99ad8d0e101b996221b8ecc2df8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":827,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[413];split=linux_only;harness=openhands;model=claude-sonnet-4-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":827,"metadata":{"avgScore":0.19774111108214873,"avg_score":0.19774111108214873,"avg_score_percent":19.77411110821487,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":242.67,"costSource":"official","cost_source":"official","harness":"openhands","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":349872550.5,"metric_id":"pass_rate","metric_role":"primary","model":"claude-sonnet-4-6","output_tokens":4376178.5,"passRate":0.09047619047619047,"pass_rate":0.09047619047619047,"pass_rate_percent":9.047619047619047,"passes":28,"runs":285,"runtime":128828.19181078624,"source_flags":[],"source_harness":"openhands","source_row_index":413,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":97,"tokens":null,"totalCostUsd":242.67,"totalDurationS":128828.19181078624,"totalInputTokens":349872550.5,"totalOutputTokens":4376178.5,"total_cost_usd":242.67,"total_input_tokens":349872550.5,"total_output_tokens":4376178.5,"total_runtime_s":128828.19181078624,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.047619047619047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-571e7981e8fef649dadd","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-149c10e907f90ff68dc74e45","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"official","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"hermes","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13174603174603172,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29ca005bffc2b99ad8d0e101b996221b8ecc2df8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":807,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[403];split=linux_only;harness=hermes;model=claude-sonnet-4-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":807,"metadata":{"avgScore":0.3203670919271656,"avg_score":0.3203670919271656,"avg_score_percent":32.036709192716565,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":436.6,"costSource":"official","cost_source":"official","harness":"hermes","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":195596773.66666666,"metric_id":"pass_rate","metric_role":"primary","model":"claude-sonnet-4-6","output_tokens":2001710.3333333333,"passRate":0.13174603174603172,"pass_rate":0.13174603174603172,"pass_rate_percent":13.174603174603172,"passes":36,"runs":248,"runtime":90692.90777800296,"source_flags":[],"source_harness":"hermes","source_row_index":403,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":96,"tokens":null,"totalCostUsd":436.6,"totalDurationS":90692.90777800296,"totalInputTokens":195596773.66666666,"totalOutputTokens":2001710.3333333333,"total_cost_usd":436.6,"total_input_tokens":195596773.66666666,"total_output_tokens":2001710.3333333333,"total_runtime_s":90692.90777800296,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.174603174603172,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c10212e6b2bd4be02f86","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f624435014a5daf8c9ca8fb5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03193666666666666,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1176,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[587];split=unlicensed/last-exam;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1176,"metadata":{"avgScore":0.03193666666666666,"avg_score":0.03193666666666666,"avg_score_percent":3.193666666666666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":92.25,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":189672323.5,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":1548284.1666666667,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":54,"runtime":126648.38385215859,"source_flags":[],"source_harness":"openclaw","source_row_index":587,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":92.25,"totalDurationS":126648.38385215859,"totalInputTokens":189672323.5,"totalOutputTokens":1548284.1666666667,"total_cost_usd":92.25,"total_input_tokens":189672323.5,"total_output_tokens":1548284.1666666667,"total_runtime_s":126648.38385215859,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.193666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-597cdf6a3105bfc3cc08","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e5d252ef34f082ce1631a0df","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2762014846817647,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":102,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[50];split=full;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":102,"metadata":{"avgScore":0.2762014846817647,"avg_score":0.2762014846817647,"avg_score_percent":27.62014846817647,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":273.05,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":511717582.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":4746342.166666667,"passRate":0.12390350877192982,"pass_rate":0.12390350877192982,"pass_rate_percent":12.390350877192983,"passes":39,"runs":275,"runtime":470844.49271820637,"source_flags":[],"source_harness":"openclaw","source_row_index":50,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":273.05,"totalDurationS":470844.49271820637,"totalInputTokens":511717582.8333333,"totalOutputTokens":4746342.166666667,"total_cost_usd":273.05,"total_input_tokens":511717582.8333333,"total_output_tokens":4746342.166666667,"total_runtime_s":470844.49271820637,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":27.62014846817647,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bbf6557e1c8bc8ecbd92","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ccd35875c5ba3d1d8615cddc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2762014846817647,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":574,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[286];split=full/overall;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":574,"metadata":{"avgScore":0.2762014846817647,"avg_score":0.2762014846817647,"avg_score_percent":27.62014846817647,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":273.05,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":511717582.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":4746342.166666667,"passRate":0.12390350877192982,"pass_rate":0.12390350877192982,"pass_rate_percent":12.390350877192983,"passes":39,"runs":275,"runtime":470844.49271820637,"source_flags":[],"source_harness":"openclaw","source_row_index":286,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":273.05,"totalDurationS":470844.49271820637,"totalInputTokens":511717582.8333333,"totalOutputTokens":4746342.166666667,"total_cost_usd":273.05,"total_input_tokens":511717582.8333333,"total_output_tokens":4746342.166666667,"total_runtime_s":470844.49271820637,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":27.62014846817647,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d85b4896fdd6a6bf0276","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c90d8921ee02ea1f10021c17","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4381171518153468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1292,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[645];split=unlicensed/near-term;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1292,"metadata":{"avgScore":0.4381171518153468,"avg_score":0.4381171518153468,"avg_score_percent":43.81171518153468,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":109.06,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":208154198.33333334,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":1909336.6666666667,"passRate":0.19900497512437812,"pass_rate":0.19900497512437812,"pass_rate_percent":19.900497512437813,"passes":28,"runs":133,"runtime":211812.18801494362,"source_flags":[],"source_harness":"openclaw","source_row_index":645,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":109.06,"totalDurationS":211812.18801494362,"totalInputTokens":208154198.33333334,"totalOutputTokens":1909336.6666666667,"total_cost_usd":109.06,"total_input_tokens":208154198.33333334,"total_output_tokens":1909336.6666666667,"total_runtime_s":211812.18801494362,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":43.81171518153468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e157985cf6b2b4874b33","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c01406e5540c455c97336a5c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24938953000000005,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1046,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[522];split=unlicensed/full-spectrum;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1046,"metadata":{"avgScore":0.24938953000000005,"avg_score":0.24938953000000005,"avg_score_percent":24.938953000000005,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":57.34,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":99504576.5,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":1380513.8333333333,"passRate":0.11,"pass_rate":0.11,"pass_rate_percent":11.0,"passes":11,"runs":91,"runtime":105415.88512140546,"source_flags":[],"source_harness":"openclaw","source_row_index":522,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":57.34,"totalDurationS":105415.88512140546,"totalInputTokens":99504576.5,"totalOutputTokens":1380513.8333333333,"total_cost_usd":57.34,"total_input_tokens":99504576.5,"total_output_tokens":1380513.8333333333,"total_runtime_s":105415.88512140546,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.938953000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7a3481c30b04e848c0ad","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b8e792151be44ba282e526af","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.010909090909090908,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":702,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[350];split=licensed;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":702,"metadata":{"avgScore":0.010909090909090908,"avg_score":0.010909090909090908,"avg_score_percent":1.0909090909090908,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":21.21,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":27131876,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":168389,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":8,"runtime":40957.625895123885,"source_flags":[],"source_harness":"openclaw","source_row_index":350,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":21.21,"totalDurationS":40957.625895123885,"totalInputTokens":27131876,"totalOutputTokens":168389,"total_cost_usd":21.21,"total_input_tokens":27131876,"total_output_tokens":168389,"total_runtime_s":40957.625895123885,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":1.0909090909090908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-21b8dd6f41fe1459996b","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-913956cc246d4f9ab7e1bc3b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2968980544087109,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":936,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[467];split=unlicensed;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":936,"metadata":{"avgScore":0.2968980544087109,"avg_score":0.2968980544087109,"avg_score_percent":29.68980544087109,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":251.83,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":484585706.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":4577953.166666667,"passRate":0.13356973995271867,"pass_rate":0.13356973995271867,"pass_rate_percent":13.356973995271867,"passes":39,"runs":267,"runtime":429886.8668230823,"source_flags":[],"source_harness":"openclaw","source_row_index":467,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":251.83,"totalDurationS":429886.8668230823,"totalInputTokens":484585706.8333333,"totalOutputTokens":4577953.166666667,"total_cost_usd":251.83,"total_input_tokens":484585706.8333333,"total_output_tokens":4577953.166666667,"total_runtime_s":429886.8668230823,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.68980544087109,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0fccbb41921e63741bec","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7978d1d20e6cdcb536fdb1f2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4381171518153468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":456,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[227];split=full/near-term;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":456,"metadata":{"avgScore":0.4381171518153468,"avg_score":0.4381171518153468,"avg_score_percent":43.81171518153468,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":109.06,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":208154198.33333334,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":1909336.6666666667,"passRate":0.19900497512437812,"pass_rate":0.19900497512437812,"pass_rate_percent":19.900497512437813,"passes":28,"runs":133,"runtime":211812.18801494362,"source_flags":[],"source_harness":"openclaw","source_row_index":227,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":109.06,"totalDurationS":211812.18801494362,"totalInputTokens":208154198.33333334,"totalOutputTokens":1909336.6666666667,"total_cost_usd":109.06,"total_input_tokens":208154198.33333334,"total_output_tokens":1909336.6666666667,"total_runtime_s":211812.18801494362,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":43.81171518153468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-17d5a9250ce50a71a154","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5a5dc2e0c4413c719222e129","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23799048181818186,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":212,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[105];split=full/full-spectrum;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":212,"metadata":{"avgScore":0.23799048181818186,"avg_score":0.23799048181818186,"avg_score_percent":23.799048181818186,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.1,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":114460972.5,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":1483905.8333333333,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":12,"runs":97,"runtime":137517.33694275937,"source_flags":[],"source_harness":"openclaw","source_row_index":105,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":68.1,"totalDurationS":137517.33694275937,"totalInputTokens":114460972.5,"totalOutputTokens":1483905.8333333333,"total_cost_usd":68.1,"total_input_tokens":114460972.5,"total_output_tokens":1483905.8333333333,"total_runtime_s":137517.33694275937,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.799048181818186,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1806ff1d261de906bc39","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4b2e2d46d37505dd0b686417","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2968980544087109,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1408,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[703];split=unlicensed/overall;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1408,"metadata":{"avgScore":0.2968980544087109,"avg_score":0.2968980544087109,"avg_score_percent":29.68980544087109,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":251.83,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":484585706.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":4577953.166666667,"passRate":0.13356973995271867,"pass_rate":0.13356973995271867,"pass_rate_percent":13.356973995271867,"passes":39,"runs":267,"runtime":429886.8668230823,"source_flags":[],"source_harness":"openclaw","source_row_index":703,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":251.83,"totalDurationS":429886.8668230823,"totalInputTokens":484585706.8333333,"totalOutputTokens":4577953.166666667,"total_cost_usd":251.83,"total_input_tokens":484585706.8333333,"total_output_tokens":4577953.166666667,"total_runtime_s":429886.8668230823,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.68980544087109,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8f08e4d4dd0ae8d737b8","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4994bfa0282ed7f170e77aa4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.30841640658848457,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":814,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[406];split=linux_only;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":814,"metadata":{"avgScore":0.30841640658848457,"avg_score":0.30841640658848457,"avg_score_percent":30.84164065884846,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":161.39,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":312819738.8333333,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":3293259.1666666665,"passRate":0.14126984126984127,"pass_rate":0.14126984126984127,"pass_rate_percent":14.126984126984127,"passes":32,"runs":213,"runtime":202301.7898809551,"source_flags":[],"source_harness":"openclaw","source_row_index":406,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":161.39,"totalDurationS":202301.7898809551,"totalInputTokens":312819738.8333333,"totalOutputTokens":3293259.1666666665,"total_cost_usd":161.39,"total_input_tokens":312819738.8333333,"total_output_tokens":3293259.1666666665,"total_runtime_s":202301.7898809551,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":30.84164065884846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5aae8e8d18a82dbff0ed","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2878b23a1287e6936c378a1e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.02521315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3091b8813ec1846c0a672c5810cb37c73c733b","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":340,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[169];split=full/last-exam;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":340,"metadata":{"avgScore":0.02521315789473684,"avg_score":0.02521315789473684,"avg_score_percent":2.521315789473684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":108.13,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":208751569.5,"metric_id":"avg_score","metric_role":"secondary","model":"deepseek-v4-pro","output_tokens":1659819.1666666667,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":59,"runtime":140834.02247361047,"source_flags":[],"source_harness":"openclaw","source_row_index":169,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":108.13,"totalDurationS":140834.02247361047,"totalInputTokens":208751569.5,"totalOutputTokens":1659819.1666666667,"total_cost_usd":108.13,"total_input_tokens":208751569.5,"total_output_tokens":1659819.1666666667,"total_runtime_s":140834.02247361047,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.521315789473684,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f994e402ea687c5bac1f","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e39e88590c45ed0f3bb4c112","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":211,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[105];split=full/full-spectrum;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":211,"metadata":{"avgScore":0.23799048181818186,"avg_score":0.23799048181818186,"avg_score_percent":23.799048181818186,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.1,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":114460972.5,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":1483905.8333333333,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":12,"runs":97,"runtime":137517.33694275937,"source_flags":[],"source_harness":"openclaw","source_row_index":105,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":68.1,"totalDurationS":137517.33694275937,"totalInputTokens":114460972.5,"totalOutputTokens":1483905.8333333333,"total_cost_usd":68.1,"total_input_tokens":114460972.5,"total_output_tokens":1483905.8333333333,"total_runtime_s":137517.33694275937,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.909090909090908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-95011447e237063d2cff","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ddd35ba0e9108c101d4c6125","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1045,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[522];split=unlicensed/full-spectrum;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1045,"metadata":{"avgScore":0.24938953000000005,"avg_score":0.24938953000000005,"avg_score_percent":24.938953000000005,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":57.34,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":99504576.5,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":1380513.8333333333,"passRate":0.11,"pass_rate":0.11,"pass_rate_percent":11.0,"passes":11,"runs":91,"runtime":105415.88512140546,"source_flags":[],"source_harness":"openclaw","source_row_index":522,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":57.34,"totalDurationS":105415.88512140546,"totalInputTokens":99504576.5,"totalOutputTokens":1380513.8333333333,"total_cost_usd":57.34,"total_input_tokens":99504576.5,"total_output_tokens":1380513.8333333333,"total_runtime_s":105415.88512140546,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b749a6ddb0bee434de98","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a442d5b09f7e1b497807bf36","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.14126984126984127,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":813,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[406];split=linux_only;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":813,"metadata":{"avgScore":0.30841640658848457,"avg_score":0.30841640658848457,"avg_score_percent":30.84164065884846,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":161.39,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":312819738.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":3293259.1666666665,"passRate":0.14126984126984127,"pass_rate":0.14126984126984127,"pass_rate_percent":14.126984126984127,"passes":32,"runs":213,"runtime":202301.7898809551,"source_flags":[],"source_harness":"openclaw","source_row_index":406,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":161.39,"totalDurationS":202301.7898809551,"totalInputTokens":312819738.8333333,"totalOutputTokens":3293259.1666666665,"total_cost_usd":161.39,"total_input_tokens":312819738.8333333,"total_output_tokens":3293259.1666666665,"total_runtime_s":202301.7898809551,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.126984126984127,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-29ec0b2311851b28cd91","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9b3b3133b79d90e399b6197f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12390350877192982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":101,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[50];split=full;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":101,"metadata":{"avgScore":0.2762014846817647,"avg_score":0.2762014846817647,"avg_score_percent":27.62014846817647,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":273.05,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":511717582.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":4746342.166666667,"passRate":0.12390350877192982,"pass_rate":0.12390350877192982,"pass_rate_percent":12.390350877192983,"passes":39,"runs":275,"runtime":470844.49271820637,"source_flags":[],"source_harness":"openclaw","source_row_index":50,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":273.05,"totalDurationS":470844.49271820637,"totalInputTokens":511717582.8333333,"totalOutputTokens":4746342.166666667,"total_cost_usd":273.05,"total_input_tokens":511717582.8333333,"total_output_tokens":4746342.166666667,"total_runtime_s":470844.49271820637,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.390350877192983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d49c0ffb4f9d99a66f3f","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-98c56184313afece11d64719","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19900497512437812,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1291,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[645];split=unlicensed/near-term;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1291,"metadata":{"avgScore":0.4381171518153468,"avg_score":0.4381171518153468,"avg_score_percent":43.81171518153468,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":109.06,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":208154198.33333334,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":1909336.6666666667,"passRate":0.19900497512437812,"pass_rate":0.19900497512437812,"pass_rate_percent":19.900497512437813,"passes":28,"runs":133,"runtime":211812.18801494362,"source_flags":[],"source_harness":"openclaw","source_row_index":645,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":109.06,"totalDurationS":211812.18801494362,"totalInputTokens":208154198.33333334,"totalOutputTokens":1909336.6666666667,"total_cost_usd":109.06,"total_input_tokens":208154198.33333334,"total_output_tokens":1909336.6666666667,"total_runtime_s":211812.18801494362,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.900497512437813,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6b71bafcc24b40e57812","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-984212527a08ee769875e408","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":339,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[169];split=full/last-exam;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":339,"metadata":{"avgScore":0.02521315789473684,"avg_score":0.02521315789473684,"avg_score_percent":2.521315789473684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":108.13,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":208751569.5,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":1659819.1666666667,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":59,"runtime":140834.02247361047,"source_flags":[],"source_harness":"openclaw","source_row_index":169,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":108.13,"totalDurationS":140834.02247361047,"totalInputTokens":208751569.5,"totalOutputTokens":1659819.1666666667,"total_cost_usd":108.13,"total_input_tokens":208751569.5,"total_output_tokens":1659819.1666666667,"total_runtime_s":140834.02247361047,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e074f2ea268d82abb996","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-727782f3fc986a819b4fee2a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13356973995271867,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":935,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[467];split=unlicensed;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":935,"metadata":{"avgScore":0.2968980544087109,"avg_score":0.2968980544087109,"avg_score_percent":29.68980544087109,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":251.83,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":484585706.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":4577953.166666667,"passRate":0.13356973995271867,"pass_rate":0.13356973995271867,"pass_rate_percent":13.356973995271867,"passes":39,"runs":267,"runtime":429886.8668230823,"source_flags":[],"source_harness":"openclaw","source_row_index":467,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":251.83,"totalDurationS":429886.8668230823,"totalInputTokens":484585706.8333333,"totalOutputTokens":4577953.166666667,"total_cost_usd":251.83,"total_input_tokens":484585706.8333333,"total_output_tokens":4577953.166666667,"total_runtime_s":429886.8668230823,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.356973995271867,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a2ee57a7b40fd1546940","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-52503a690a5794ff4021191c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1175,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[587];split=unlicensed/last-exam;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1175,"metadata":{"avgScore":0.03193666666666666,"avg_score":0.03193666666666666,"avg_score_percent":3.193666666666666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":92.25,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":189672323.5,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":1548284.1666666667,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":54,"runtime":126648.38385215859,"source_flags":[],"source_harness":"openclaw","source_row_index":587,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":92.25,"totalDurationS":126648.38385215859,"totalInputTokens":189672323.5,"totalOutputTokens":1548284.1666666667,"total_cost_usd":92.25,"total_input_tokens":189672323.5,"total_output_tokens":1548284.1666666667,"total_runtime_s":126648.38385215859,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e960b6dd894861039f40","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-38d17b92c542d3d820b14837","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13356973995271867,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1407,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[703];split=unlicensed/overall;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1407,"metadata":{"avgScore":0.2968980544087109,"avg_score":0.2968980544087109,"avg_score_percent":29.68980544087109,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":251.83,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":484585706.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":4577953.166666667,"passRate":0.13356973995271867,"pass_rate":0.13356973995271867,"pass_rate_percent":13.356973995271867,"passes":39,"runs":267,"runtime":429886.8668230823,"source_flags":[],"source_harness":"openclaw","source_row_index":703,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":251.83,"totalDurationS":429886.8668230823,"totalInputTokens":484585706.8333333,"totalOutputTokens":4577953.166666667,"total_cost_usd":251.83,"total_input_tokens":484585706.8333333,"total_output_tokens":4577953.166666667,"total_runtime_s":429886.8668230823,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.356973995271867,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-16c1477e0d58f9e24584","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3106487ce24c6e5bbae9b2b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19900497512437812,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":455,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[227];split=full/near-term;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":455,"metadata":{"avgScore":0.4381171518153468,"avg_score":0.4381171518153468,"avg_score_percent":43.81171518153468,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":109.06,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":208154198.33333334,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":1909336.6666666667,"passRate":0.19900497512437812,"pass_rate":0.19900497512437812,"pass_rate_percent":19.900497512437813,"passes":28,"runs":133,"runtime":211812.18801494362,"source_flags":[],"source_harness":"openclaw","source_row_index":227,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":109.06,"totalDurationS":211812.18801494362,"totalInputTokens":208154198.33333334,"totalOutputTokens":1909336.6666666667,"total_cost_usd":109.06,"total_input_tokens":208154198.33333334,"total_output_tokens":1909336.6666666667,"total_runtime_s":211812.18801494362,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.900497512437813,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6957178f7a009e4149b0","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1bdfe5862e023d438f65fdcc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12390350877192982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":573,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[286];split=full/overall;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":573,"metadata":{"avgScore":0.2762014846817647,"avg_score":0.2762014846817647,"avg_score_percent":27.62014846817647,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":273.05,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":511717582.8333333,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":4746342.166666667,"passRate":0.12390350877192982,"pass_rate":0.12390350877192982,"pass_rate_percent":12.390350877192983,"passes":39,"runs":275,"runtime":470844.49271820637,"source_flags":[],"source_harness":"openclaw","source_row_index":286,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":273.05,"totalDurationS":470844.49271820637,"totalInputTokens":511717582.8333333,"totalOutputTokens":4746342.166666667,"total_cost_usd":273.05,"total_input_tokens":511717582.8333333,"total_output_tokens":4746342.166666667,"total_runtime_s":470844.49271820637,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.390350877192983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8ddb8fb28ccdd57fd975","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0594c26811c2a2f4558a5968","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57fe5ce354d77b0d97cf4cf0930bcfe7178cf9b1","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":701,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[350];split=licensed;harness=openclaw;model=deepseek-v4-pro;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":701,"metadata":{"avgScore":0.010909090909090908,"avg_score":0.010909090909090908,"avg_score_percent":1.0909090909090908,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":21.21,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":27131876,"metric_id":"pass_rate","metric_role":"primary","model":"deepseek-v4-pro","output_tokens":168389,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":8,"runtime":40957.625895123885,"source_flags":[],"source_harness":"openclaw","source_row_index":350,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":21.21,"totalDurationS":40957.625895123885,"totalInputTokens":27131876,"totalOutputTokens":168389,"total_cost_usd":21.21,"total_input_tokens":27131876,"total_output_tokens":168389,"total_runtime_s":40957.625895123885,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8e957ad20aaf0ba33641","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e7853f957319839f915f625e","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3480542170776246,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":922,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[460];split=unlicensed;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":922,"metadata":{"avgScore":0.3480542170776246,"avg_score":0.3480542170776246,"avg_score_percent":34.805421707762456,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1942.97,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1159533595,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3389319,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":135,"runtime":317055,"source_flags":[],"source_harness":"gemini_cli","source_row_index":460,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1942.97,"totalDurationS":317055,"totalInputTokens":1159533595,"totalOutputTokens":3389319,"total_cost_usd":1942.97,"total_input_tokens":1159533595,"total_output_tokens":3389319,"total_runtime_s":317055,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":34.805421707762456,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-db350f5c57d6f3c94e62","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e69af8736b1cdfc5180e81b5","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.31130806689351526,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1404,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[701];split=unlicensed/overall;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1404,"metadata":{"avgScore":0.31130806689351526,"avg_score":0.31130806689351526,"avg_score_percent":31.130806689351527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2624.2,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2913806051,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3449466,"passRate":0.1524822695035461,"pass_rate":0.1524822695035461,"pass_rate_percent":15.24822695035461,"passes":22,"runs":142,"runtime":449865.6185256547,"source_flags":[],"source_harness":"openclaw","source_row_index":701,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":2624.2,"totalDurationS":449865.6185256547,"totalInputTokens":2913806051,"totalOutputTokens":3449466,"total_cost_usd":2624.2,"total_input_tokens":2913806051,"total_output_tokens":3449466,"total_runtime_s":449865.6185256547,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.130806689351527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f21ebfd0caa40e7d1509","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ca78c576d48d33ec655329fb","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3270766092627965,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":558,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[278];split=full/overall;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":558,"metadata":{"avgScore":0.3270766092627965,"avg_score":0.3270766092627965,"avg_score_percent":32.70766092627965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2018.05,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1212565963,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3542216,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":142,"runtime":332449,"source_flags":[],"source_harness":"gemini_cli","source_row_index":278,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":2018.05,"totalDurationS":332449,"totalInputTokens":1212565963,"totalOutputTokens":3542216,"total_cost_usd":2018.05,"total_input_tokens":1212565963,"total_output_tokens":3542216,"total_runtime_s":332449,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.70766092627965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b8a7fba19103e1500ad0","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c97e533c5001e7f3961d5d37","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4954053099301343,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":444,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[221];split=full/near-term;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":444,"metadata":{"avgScore":0.4954053099301343,"avg_score":0.4954053099301343,"avg_score_percent":49.540530993013434,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":574.73,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":616736935.5,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":938928.5,"passRate":0.26119402985074625,"pass_rate":0.26119402985074625,"pass_rate_percent":26.119402985074625,"passes":18,"runs":67,"runtime":175033.16224954824,"source_flags":[],"source_harness":"openclaw","source_row_index":221,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":574.73,"totalDurationS":175033.16224954824,"totalInputTokens":616736935.5,"totalOutputTokens":938928.5,"total_cost_usd":574.73,"total_input_tokens":616736935.5,"total_output_tokens":938928.5,"total_runtime_s":175033.16224954824,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":49.540530993013434,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2646dcee8fb7741ef899","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-be239a7475ed1697ce8c9336","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3717262877253931,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":794,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[396];split=linux_only;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":794,"metadata":{"avgScore":0.3717262877253931,"avg_score":0.3717262877253931,"avg_score_percent":37.17262877253931,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1009.61,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":450754861,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":2390671,"passRate":0.18095238095238095,"pass_rate":0.18095238095238095,"pass_rate_percent":18.095238095238095,"passes":19,"runs":101,"runtime":188815,"source_flags":[],"source_harness":"gemini_cli","source_row_index":396,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":1009.61,"totalDurationS":188815,"totalInputTokens":450754861,"totalOutputTokens":2390671,"total_cost_usd":1009.61,"total_input_tokens":450754861,"total_output_tokens":2390671,"total_runtime_s":188815,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.17262877253931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-22373d5f7f88fbf9e4fa","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bcb00d1d18ada9dd7083149c","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.31130806689351526,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":932,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[465];split=unlicensed;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":932,"metadata":{"avgScore":0.31130806689351526,"avg_score":0.31130806689351526,"avg_score_percent":31.130806689351527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2624.2,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2913806051,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3449466,"passRate":0.1524822695035461,"pass_rate":0.1524822695035461,"pass_rate_percent":15.24822695035461,"passes":22,"runs":142,"runtime":449865.6185256547,"source_flags":[],"source_harness":"openclaw","source_row_index":465,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":2624.2,"totalDurationS":449865.6185256547,"totalInputTokens":2913806051,"totalOutputTokens":3449466,"total_cost_usd":2624.2,"total_input_tokens":2913806051,"total_output_tokens":3449466,"total_runtime_s":449865.6185256547,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.130806689351527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c2912ef1c3eb6381aefc","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b60dd410b21cd44aec5e5c04","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2927265620525372,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":96,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[47];split=full;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":96,"metadata":{"avgScore":0.2927265620525372,"avg_score":0.2927265620525372,"avg_score_percent":29.27265620525372,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2968.62,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":3369212141,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3653426,"passRate":0.14144736842105263,"pass_rate":0.14144736842105263,"pass_rate_percent":14.144736842105262,"passes":22,"runs":149,"runtime":529548.719956005,"source_flags":[],"source_harness":"openclaw","source_row_index":47,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":2968.62,"totalDurationS":529548.719956005,"totalInputTokens":3369212141,"totalOutputTokens":3653426,"total_cost_usd":2968.62,"total_input_tokens":3369212141,"total_output_tokens":3653426,"total_runtime_s":529548.719956005,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.27265620525372,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b53319ea150df02aad20","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b42cd2aa17066a153978ee64","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3270766092627965,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":86,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[42];split=full;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":86,"metadata":{"avgScore":0.3270766092627965,"avg_score":0.3270766092627965,"avg_score_percent":32.70766092627965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2018.05,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1212565963,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3542216,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":142,"runtime":332449,"source_flags":[],"source_harness":"gemini_cli","source_row_index":42,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":2018.05,"totalDurationS":332449,"totalInputTokens":1212565963,"totalOutputTokens":3542216,"total_cost_usd":2018.05,"total_input_tokens":1212565963,"total_output_tokens":3542216,"total_runtime_s":332449,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.70766092627965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-667030b5926df9743c3e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a79037db55cf6334f953b0d2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03868333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1172,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[585];split=unlicensed/last-exam;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1172,"metadata":{"avgScore":0.03868333333333333,"avg_score":0.03868333333333333,"avg_score_percent":3.868333333333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":897.91,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":959876973.5,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1051546.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":107122.25937282527,"source_flags":[],"source_harness":"openclaw","source_row_index":585,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":897.91,"totalDurationS":107122.25937282527,"totalInputTokens":959876973.5,"totalOutputTokens":1051546.5,"total_cost_usd":897.91,"total_input_tokens":959876973.5,"total_output_tokens":1051546.5,"total_runtime_s":107122.25937282527,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.868333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-18bd6b3998f86c475d1c","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a6d01db319fe00efbf44be86","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.009123684210526317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":352,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[175];split=full/last-exam;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":352,"metadata":{"avgScore":0.009123684210526317,"avg_score":0.009123684210526317,"avg_score_percent":0.9123684210526317,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":732.51,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":497315638,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":821016,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":144134,"source_flags":[],"source_harness":"gemini_cli","source_row_index":175,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":30,"tokens":null,"totalCostUsd":732.51,"totalDurationS":144134,"totalInputTokens":497315638,"totalOutputTokens":821016,"total_cost_usd":732.51,"total_input_tokens":497315638,"total_output_tokens":821016,"total_runtime_s":144134,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.9123684210526317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-43965c94d773fd686a25","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a5e5058ff7dd5ac49be97caf","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5494600574739001,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":440,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[219];split=full/near-term;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":440,"metadata":{"avgScore":0.5494600574739001,"avg_score":0.5494600574739001,"avg_score_percent":54.94600574739002,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":341.66,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":239360406,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1099196,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":65,"runtime":99601,"source_flags":[],"source_harness":"gemini_cli","source_row_index":219,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":341.66,"totalDurationS":99601,"totalInputTokens":239360406,"totalOutputTokens":1099196,"total_cost_usd":341.66,"total_input_tokens":239360406,"total_output_tokens":1099196,"total_runtime_s":99601,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":54.94600574739002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2bb1d860e770c77e808a","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7beb12704b2b91f7d347be94","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.011556666666666666,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1188,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[593];split=unlicensed/last-exam;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1188,"metadata":{"avgScore":0.011556666666666666,"avg_score":0.011556666666666666,"avg_score_percent":1.1556666666666666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":729.11,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":496016340,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":813079,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":26,"runtime":141875,"source_flags":[],"source_harness":"gemini_cli","source_row_index":593,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":729.11,"totalDurationS":141875,"totalInputTokens":496016340,"totalOutputTokens":813079,"total_cost_usd":729.11,"total_input_tokens":496016340,"total_output_tokens":813079,"total_runtime_s":141875,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":1.1556666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6255a7b552b8b1bed87f","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7671e5368715653164119a29","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.25745041514387507,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1044,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[521];split=unlicensed/full-spectrum;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1044,"metadata":{"avgScore":0.25745041514387507,"avg_score":0.25745041514387507,"avg_score_percent":25.745041514387506,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":888.72,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":429544179,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1577347,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":49,"runtime":78150,"source_flags":[],"source_harness":"gemini_cli","source_row_index":521,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":888.72,"totalDurationS":78150,"totalInputTokens":429544179,"totalOutputTokens":1577347,"total_cost_usd":888.72,"total_input_tokens":429544179,"total_output_tokens":1577347,"total_runtime_s":78150,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":25.745041514387506,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-40cd2c38d1cae3c2ef34","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-728389d5d6295a092647ab7d","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.05454545454545454,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":688,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[343];split=licensed;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":688,"metadata":{"avgScore":0.05454545454545454,"avg_score":0.05454545454545454,"avg_score_percent":5.454545454545454,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":344.43,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":455406090,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":203960,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":79683.10143035022,"source_flags":[],"source_harness":"openclaw","source_row_index":343,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":344.43,"totalDurationS":79683.10143035022,"totalInputTokens":455406090,"totalOutputTokens":203960,"total_cost_usd":344.43,"total_input_tokens":455406090,"total_output_tokens":203960,"total_runtime_s":79683.10143035022,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":5.454545454545454,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5922c622a267b7614db0","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6c5afd12ef0583d93fa4d856","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22743563333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1054,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[526];split=unlicensed/full-spectrum;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1054,"metadata":{"avgScore":0.22743563333333333,"avg_score":0.22743563333333333,"avg_score_percent":22.743563333333334,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1191.3,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1373512411,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1585963,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":5,"runs":51,"runtime":183302.91810718234,"source_flags":[],"source_harness":"openclaw","source_row_index":526,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":47,"tokens":null,"totalCostUsd":1191.3,"totalDurationS":183302.91810718234,"totalInputTokens":1373512411,"totalOutputTokens":1585963,"total_cost_usd":1191.3,"total_input_tokens":1373512411,"total_output_tokens":1585963,"total_runtime_s":183302.91810718234,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.743563333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-167368ccbbda58d657dd","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-68f98fd60f091a336e1f8687","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2927265620525372,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":568,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[283];split=full/overall;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":568,"metadata":{"avgScore":0.2927265620525372,"avg_score":0.2927265620525372,"avg_score_percent":29.27265620525372,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2968.62,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":3369212141,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3653426,"passRate":0.14144736842105263,"pass_rate":0.14144736842105263,"pass_rate_percent":14.144736842105262,"passes":22,"runs":149,"runtime":529548.719956005,"source_flags":[],"source_harness":"openclaw","source_row_index":283,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":2968.62,"totalDurationS":529548.719956005,"totalInputTokens":3369212141,"totalOutputTokens":3653426,"total_cost_usd":2968.62,"total_input_tokens":3369212141,"total_output_tokens":3653426,"total_runtime_s":529548.719956005,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.27265620525372,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-81616126bb20c9ced99e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-48852245cfd4101a46429e8d","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4954053099301343,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1280,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[639];split=unlicensed/near-term;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1280,"metadata":{"avgScore":0.4954053099301343,"avg_score":0.4954053099301343,"avg_score_percent":49.540530993013434,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":574.73,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":616736935.5,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":938928.5,"passRate":0.26119402985074625,"pass_rate":0.26119402985074625,"pass_rate_percent":26.119402985074625,"passes":18,"runs":67,"runtime":175033.16224954824,"source_flags":[],"source_harness":"openclaw","source_row_index":639,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":574.73,"totalDurationS":175033.16224954824,"totalInputTokens":616736935.5,"totalOutputTokens":938928.5,"total_cost_usd":574.73,"total_input_tokens":616736935.5,"total_output_tokens":938928.5,"total_runtime_s":175033.16224954824,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":49.540530993013434,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7970fc1650271b4132fc","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3483bc5f8f49a8c302bf61f2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23585057575757576,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":214,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[106];split=full/full-spectrum;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":214,"metadata":{"avgScore":0.23585057575757576,"avg_score":0.23585057575757576,"avg_score_percent":23.585057575757578,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1391.91,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1636895147,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1744287,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":56,"runtime":226023.81774288748,"source_flags":[],"source_harness":"openclaw","source_row_index":106,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":52,"tokens":null,"totalCostUsd":1391.91,"totalDurationS":226023.81774288748,"totalInputTokens":1636895147,"totalOutputTokens":1744287,"total_cost_usd":1391.91,"total_input_tokens":1636895147,"total_output_tokens":1744287,"total_runtime_s":226023.81774288748,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.585057575757578,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ef66adff4f5695a641a0","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2c5c4014a7ae58adedc3e59e","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3480542170776246,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1394,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[696];split=unlicensed/overall;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1394,"metadata":{"avgScore":0.3480542170776246,"avg_score":0.3480542170776246,"avg_score_percent":34.805421707762456,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1942.97,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1159533595,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":3389319,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":135,"runtime":317055,"source_flags":[],"source_harness":"gemini_cli","source_row_index":696,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1942.97,"totalDurationS":317055,"totalInputTokens":1159533595,"totalOutputTokens":3389319,"total_cost_usd":1942.97,"total_input_tokens":1159533595,"total_output_tokens":3389319,"total_runtime_s":317055,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":34.805421707762456,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-032217abba924f8eb2e6","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-246767e928c6b7258d12fed4","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5494600574739001,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1276,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[637];split=unlicensed/near-term;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1276,"metadata":{"avgScore":0.5494600574739001,"avg_score":0.5494600574739001,"avg_score_percent":54.94600574739002,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":341.66,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":239360406,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1099196,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":65,"runtime":99601,"source_flags":[],"source_harness":"gemini_cli","source_row_index":637,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":341.66,"totalDurationS":99601,"totalInputTokens":239360406,"totalOutputTokens":1099196,"total_cost_usd":341.66,"total_input_tokens":239360406,"total_output_tokens":1099196,"total_runtime_s":99601,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":54.94600574739002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fc2d78cd0805b9065410","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-235a970e2ff493fe249b1a62","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.05818181818181818,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":686,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[342];split=licensed;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":686,"metadata":{"avgScore":0.05818181818181818,"avg_score":0.05818181818181818,"avg_score_percent":5.818181818181818,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":75.08,"costSource":"db+adj","cost_source":"db+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":53032368,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":152897,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":15394,"source_flags":[],"source_harness":"gemini_cli","source_row_index":342,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":75.08,"totalDurationS":15394,"totalInputTokens":53032368,"totalOutputTokens":152897,"total_cost_usd":75.08,"total_input_tokens":53032368,"total_output_tokens":152897,"total_runtime_s":15394,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":5.818181818181818,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9bc0cfafd3221378b305","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2132661f65104389aa66e3df","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.32433446158178597,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":806,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[402];split=linux_only;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":806,"metadata":{"avgScore":0.32433446158178597,"avg_score":0.32433446158178597,"avg_score_percent":32.4334461581786,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1303.46,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1367870276,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":2349345,"passRate":0.15714285714285714,"pass_rate":0.15714285714285714,"pass_rate_percent":15.714285714285714,"passes":17,"runs":108,"runtime":238983.29079494995,"source_flags":[],"source_harness":"openclaw","source_row_index":402,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":100,"tokens":null,"totalCostUsd":1303.46,"totalDurationS":238983.29079494995,"totalInputTokens":1367870276,"totalOutputTokens":2349345,"total_cost_usd":1303.46,"total_input_tokens":1367870276,"total_output_tokens":2349345,"total_runtime_s":238983.29079494995,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.4334461581786,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8a190d21de901a95c2f1","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-205f79d357f9c2cdafc4bcc3","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2638640137671591,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":206,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[102];split=full/full-spectrum;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":206,"metadata":{"avgScore":0.2638640137671591,"avg_score":0.2638640137671591,"avg_score_percent":26.386401376715913,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":961.76,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":481651301,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1728321,"passRate":0.12727272727272726,"pass_rate":0.12727272727272726,"pass_rate_percent":12.727272727272727,"passes":7,"runs":54,"runtime":92042,"source_flags":[],"source_harness":"gemini_cli","source_row_index":102,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":961.76,"totalDurationS":92042,"totalInputTokens":481651301,"totalOutputTokens":1728321,"total_cost_usd":961.76,"total_input_tokens":481651301,"total_output_tokens":1728321,"total_runtime_s":92042,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.386401376715913,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c773e9f055323b0fc3a4","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-05d758475e58520b0d258242","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03053947368421053,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c32193b18f4324b3f16e17eede85d278e3f9a6b7","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":338,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[168];split=full/last-exam;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":338,"metadata":{"avgScore":0.03053947368421053,"avg_score":0.03053947368421053,"avg_score_percent":3.0539473684210527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1200.21,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1379338229.5,"metric_id":"avg_score","metric_role":"secondary","model":"gemini-3-1-pro-preview","output_tokens":1139995.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":34,"runtime":162523.19375644717,"source_flags":[],"source_harness":"openclaw","source_row_index":168,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":1200.21,"totalDurationS":162523.19375644717,"totalInputTokens":1379338229.5,"totalOutputTokens":1139995.5,"total_cost_usd":1200.21,"total_input_tokens":1379338229.5,"total_output_tokens":1139995.5,"total_runtime_s":162523.19375644717,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.0539473684210527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1623f416969ca5447e4e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f9d68efcb1210b8e06d00e25","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.16447368421052633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":85,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[42];split=full;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":85,"metadata":{"avgScore":0.3270766092627965,"avg_score":0.3270766092627965,"avg_score_percent":32.70766092627965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2018.05,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1212565963,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3542216,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":142,"runtime":332449,"source_flags":[],"source_harness":"gemini_cli","source_row_index":42,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":2018.05,"totalDurationS":332449,"totalInputTokens":1212565963,"totalOutputTokens":3542216,"total_cost_usd":2018.05,"total_input_tokens":1212565963,"total_output_tokens":3542216,"total_runtime_s":332449,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":16.447368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d2e11cd7442ca32160d2","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f452cb53123922e2955d92b2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1275,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[637];split=unlicensed/near-term;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1275,"metadata":{"avgScore":0.5494600574739001,"avg_score":0.5494600574739001,"avg_score_percent":54.94600574739002,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":341.66,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":239360406,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1099196,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":65,"runtime":99601,"source_flags":[],"source_harness":"gemini_cli","source_row_index":637,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":341.66,"totalDurationS":99601,"totalInputTokens":239360406,"totalOutputTokens":1099196,"total_cost_usd":341.66,"total_input_tokens":239360406,"total_output_tokens":1099196,"total_runtime_s":99601,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-25e34457e4420f3ee592","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e7f7018558abe22b973f459e","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18095238095238095,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":793,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[396];split=linux_only;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":793,"metadata":{"avgScore":0.3717262877253931,"avg_score":0.3717262877253931,"avg_score_percent":37.17262877253931,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1009.61,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":450754861,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":2390671,"passRate":0.18095238095238095,"pass_rate":0.18095238095238095,"pass_rate_percent":18.095238095238095,"passes":19,"runs":101,"runtime":188815,"source_flags":[],"source_harness":"gemini_cli","source_row_index":396,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":1009.61,"totalDurationS":188815,"totalInputTokens":450754861,"totalOutputTokens":2390671,"total_cost_usd":1009.61,"total_input_tokens":450754861,"total_output_tokens":2390671,"total_runtime_s":188815,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.095238095238095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8eec506c4531e7861e1","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e63f1f538bea6cfda8021577","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12727272727272726,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":205,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[102];split=full/full-spectrum;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":205,"metadata":{"avgScore":0.2638640137671591,"avg_score":0.2638640137671591,"avg_score_percent":26.386401376715913,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":961.76,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":481651301,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1728321,"passRate":0.12727272727272726,"pass_rate":0.12727272727272726,"pass_rate_percent":12.727272727272727,"passes":7,"runs":54,"runtime":92042,"source_flags":[],"source_harness":"gemini_cli","source_row_index":102,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":961.76,"totalDurationS":92042,"totalInputTokens":481651301,"totalOutputTokens":1728321,"total_cost_usd":961.76,"total_input_tokens":481651301,"total_output_tokens":1728321,"total_runtime_s":92042,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.727272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ad88cd666462ebeba839","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e076ef161b65e1517455b0f4","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26119402985074625,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1279,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[639];split=unlicensed/near-term;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1279,"metadata":{"avgScore":0.4954053099301343,"avg_score":0.4954053099301343,"avg_score_percent":49.540530993013434,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":574.73,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":616736935.5,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":938928.5,"passRate":0.26119402985074625,"pass_rate":0.26119402985074625,"pass_rate_percent":26.119402985074625,"passes":18,"runs":67,"runtime":175033.16224954824,"source_flags":[],"source_harness":"openclaw","source_row_index":639,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":574.73,"totalDurationS":175033.16224954824,"totalInputTokens":616736935.5,"totalOutputTokens":938928.5,"total_cost_usd":574.73,"total_input_tokens":616736935.5,"total_output_tokens":938928.5,"total_runtime_s":175033.16224954824,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.119402985074625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8690d34f8f18e39b7504","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e057ff66bda5eab5b440f0ea","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.14144736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":95,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[47];split=full;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":95,"metadata":{"avgScore":0.2927265620525372,"avg_score":0.2927265620525372,"avg_score_percent":29.27265620525372,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2968.62,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":3369212141,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3653426,"passRate":0.14144736842105263,"pass_rate":0.14144736842105263,"pass_rate_percent":14.144736842105262,"passes":22,"runs":149,"runtime":529548.719956005,"source_flags":[],"source_harness":"openclaw","source_row_index":47,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":2968.62,"totalDurationS":529548.719956005,"totalInputTokens":3369212141,"totalOutputTokens":3653426,"total_cost_usd":2968.62,"total_input_tokens":3369212141,"total_output_tokens":3653426,"total_runtime_s":529548.719956005,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.144736842105262,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-36f1dc79289df4b7117e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c5d6b421f86639a4daec1ac1","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":337,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[168];split=full/last-exam;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":337,"metadata":{"avgScore":0.03053947368421053,"avg_score":0.03053947368421053,"avg_score_percent":3.0539473684210527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1200.21,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1379338229.5,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1139995.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":34,"runtime":162523.19375644717,"source_flags":[],"source_harness":"openclaw","source_row_index":168,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":1200.21,"totalDurationS":162523.19375644717,"totalInputTokens":1379338229.5,"totalOutputTokens":1139995.5,"total_cost_usd":1200.21,"total_input_tokens":1379338229.5,"total_output_tokens":1139995.5,"total_runtime_s":162523.19375644717,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6b439d02f22c07afeb14","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c14214b235a37059a9d56988","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1524822695035461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1403,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[701];split=unlicensed/overall;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1403,"metadata":{"avgScore":0.31130806689351526,"avg_score":0.31130806689351526,"avg_score_percent":31.130806689351527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2624.2,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2913806051,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3449466,"passRate":0.1524822695035461,"pass_rate":0.1524822695035461,"pass_rate_percent":15.24822695035461,"passes":22,"runs":142,"runtime":449865.6185256547,"source_flags":[],"source_harness":"openclaw","source_row_index":701,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":2624.2,"totalDurationS":449865.6185256547,"totalInputTokens":2913806051,"totalOutputTokens":3449466,"total_cost_usd":2624.2,"total_input_tokens":2913806051,"total_output_tokens":3449466,"total_runtime_s":449865.6185256547,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.24822695035461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8b9f9512c02e5d18ad0d","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b75bd0668a154a011eef633f","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26119402985074625,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":443,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[221];split=full/near-term;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":443,"metadata":{"avgScore":0.4954053099301343,"avg_score":0.4954053099301343,"avg_score_percent":49.540530993013434,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":574.73,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":616736935.5,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":938928.5,"passRate":0.26119402985074625,"pass_rate":0.26119402985074625,"pass_rate_percent":26.119402985074625,"passes":18,"runs":67,"runtime":175033.16224954824,"source_flags":[],"source_harness":"openclaw","source_row_index":221,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":574.73,"totalDurationS":175033.16224954824,"totalInputTokens":616736935.5,"totalOutputTokens":938928.5,"total_cost_usd":574.73,"total_input_tokens":616736935.5,"total_output_tokens":938928.5,"total_runtime_s":175033.16224954824,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.119402985074625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7b621797fd7dd4c89c2c","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b0b4584bf655643036ddc71f","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":685,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[342];split=licensed;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":685,"metadata":{"avgScore":0.05818181818181818,"avg_score":0.05818181818181818,"avg_score_percent":5.818181818181818,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":75.08,"costSource":"db+adj","cost_source":"db+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":53032368,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":152897,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":15394,"source_flags":[],"source_harness":"gemini_cli","source_row_index":342,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":75.08,"totalDurationS":15394,"totalInputTokens":53032368,"totalOutputTokens":152897,"total_cost_usd":75.08,"total_input_tokens":53032368,"total_output_tokens":152897,"total_runtime_s":15394,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bd6f6b75b4aec823f874","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-839cbfc491754402bab38cad","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1773049645390071,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1393,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[696];split=unlicensed/overall;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1393,"metadata":{"avgScore":0.3480542170776246,"avg_score":0.3480542170776246,"avg_score_percent":34.805421707762456,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1942.97,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1159533595,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3389319,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":135,"runtime":317055,"source_flags":[],"source_harness":"gemini_cli","source_row_index":696,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1942.97,"totalDurationS":317055,"totalInputTokens":1159533595,"totalOutputTokens":3389319,"total_cost_usd":1942.97,"total_input_tokens":1159533595,"total_output_tokens":3389319,"total_runtime_s":317055,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.73049645390071,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8f22d642f4fb9186f003","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-829f95511808de24fd62f02c","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1187,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[593];split=unlicensed/last-exam;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1187,"metadata":{"avgScore":0.011556666666666666,"avg_score":0.011556666666666666,"avg_score_percent":1.1556666666666666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":729.11,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":496016340,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":813079,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":26,"runtime":141875,"source_flags":[],"source_harness":"gemini_cli","source_row_index":593,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":729.11,"totalDurationS":141875,"totalInputTokens":496016340,"totalOutputTokens":813079,"total_cost_usd":729.11,"total_input_tokens":496016340,"total_output_tokens":813079,"total_runtime_s":141875,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8842fec2a4a9714c7614","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7842baf0bc705093b76183b1","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1043,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[521];split=unlicensed/full-spectrum;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1043,"metadata":{"avgScore":0.25745041514387507,"avg_score":0.25745041514387507,"avg_score_percent":25.745041514387506,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":888.72,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":429544179,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1577347,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":49,"runtime":78150,"source_flags":[],"source_harness":"gemini_cli","source_row_index":521,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":888.72,"totalDurationS":78150,"totalInputTokens":429544179,"totalOutputTokens":1577347,"total_cost_usd":888.72,"total_input_tokens":429544179,"total_output_tokens":1577347,"total_runtime_s":78150,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e91b0de38b9bf3c64df6","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-72cf7bef3a379c5672e19904","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2835820895522388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":439,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[219];split=full/near-term;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":439,"metadata":{"avgScore":0.5494600574739001,"avg_score":0.5494600574739001,"avg_score_percent":54.94600574739002,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":341.66,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":239360406,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1099196,"passRate":0.2835820895522388,"pass_rate":0.2835820895522388,"pass_rate_percent":28.35820895522388,"passes":19,"runs":65,"runtime":99601,"source_flags":[],"source_harness":"gemini_cli","source_row_index":219,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":341.66,"totalDurationS":99601,"totalInputTokens":239360406,"totalOutputTokens":1099196,"total_cost_usd":341.66,"total_input_tokens":239360406,"total_output_tokens":1099196,"total_runtime_s":99601,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.35820895522388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ecd59e84995b2a892a09","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6f6079110983475f228c5d80","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":213,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[106];split=full/full-spectrum;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":213,"metadata":{"avgScore":0.23585057575757576,"avg_score":0.23585057575757576,"avg_score_percent":23.585057575757578,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1391.91,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1636895147,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1744287,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":56,"runtime":226023.81774288748,"source_flags":[],"source_harness":"openclaw","source_row_index":106,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":52,"tokens":null,"totalCostUsd":1391.91,"totalDurationS":226023.81774288748,"totalInputTokens":1636895147,"totalOutputTokens":1744287,"total_cost_usd":1391.91,"total_input_tokens":1636895147,"total_output_tokens":1744287,"total_runtime_s":226023.81774288748,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.909090909090908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a0d12fb4c7ea7c96ac31","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-62227228620041cc65b504ac","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.16447368421052633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":557,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[278];split=full/overall;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":557,"metadata":{"avgScore":0.3270766092627965,"avg_score":0.3270766092627965,"avg_score_percent":32.70766092627965,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2018.05,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1212565963,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3542216,"passRate":0.16447368421052633,"pass_rate":0.16447368421052633,"pass_rate_percent":16.447368421052634,"passes":25,"runs":142,"runtime":332449,"source_flags":[],"source_harness":"gemini_cli","source_row_index":278,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":2018.05,"totalDurationS":332449,"totalInputTokens":1212565963,"totalOutputTokens":3542216,"total_cost_usd":2018.05,"total_input_tokens":1212565963,"total_output_tokens":3542216,"total_runtime_s":332449,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":16.447368421052634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b8d479d4f527d8a67207","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5dee03e5e25d74f54e377b3e","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":351,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[175];split=full/last-exam;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":351,"metadata":{"avgScore":0.009123684210526317,"avg_score":0.009123684210526317,"avg_score_percent":0.9123684210526317,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":732.51,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":497315638,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":821016,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":144134,"source_flags":[],"source_harness":"gemini_cli","source_row_index":175,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":30,"tokens":null,"totalCostUsd":732.51,"totalDurationS":144134,"totalInputTokens":497315638,"totalOutputTokens":821016,"total_cost_usd":732.51,"total_input_tokens":497315638,"total_output_tokens":821016,"total_runtime_s":144134,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6fdff421c62793a82833","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5118ee368b8df0edf000d6ad","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"gemini_cli","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1773049645390071,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":921,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[460];split=unlicensed;harness=gemini_cli;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":921,"metadata":{"avgScore":0.3480542170776246,"avg_score":0.3480542170776246,"avg_score_percent":34.805421707762456,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1942.97,"costSource":"db+list+adj","cost_source":"db+list+adj","harness":"gemini_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1159533595,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3389319,"passRate":0.1773049645390071,"pass_rate":0.1773049645390071,"pass_rate_percent":17.73049645390071,"passes":25,"runs":135,"runtime":317055,"source_flags":[],"source_harness":"gemini_cli","source_row_index":460,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":1942.97,"totalDurationS":317055,"totalInputTokens":1159533595,"totalOutputTokens":3389319,"total_cost_usd":1942.97,"total_input_tokens":1159533595,"total_output_tokens":3389319,"total_runtime_s":317055,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.73049645390071,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a6d605df684da657015","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4d1bdde6802ee352942329d9","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1171,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[585];split=unlicensed/last-exam;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1171,"metadata":{"avgScore":0.03868333333333333,"avg_score":0.03868333333333333,"avg_score_percent":3.868333333333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":897.91,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":959876973.5,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1051546.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":107122.25937282527,"source_flags":[],"source_harness":"openclaw","source_row_index":585,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":897.91,"totalDurationS":107122.25937282527,"totalInputTokens":959876973.5,"totalOutputTokens":1051546.5,"total_cost_usd":897.91,"total_input_tokens":959876973.5,"total_output_tokens":1051546.5,"total_runtime_s":107122.25937282527,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-222e410caa7ffd37300d","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2ffbe01a995441bc468d87f4","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.15714285714285714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":805,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[402];split=linux_only;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":805,"metadata":{"avgScore":0.32433446158178597,"avg_score":0.32433446158178597,"avg_score_percent":32.4334461581786,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1303.46,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1367870276,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":2349345,"passRate":0.15714285714285714,"pass_rate":0.15714285714285714,"pass_rate_percent":15.714285714285714,"passes":17,"runs":108,"runtime":238983.29079494995,"source_flags":[],"source_harness":"openclaw","source_row_index":402,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":100,"tokens":null,"totalCostUsd":1303.46,"totalDurationS":238983.29079494995,"totalInputTokens":1367870276,"totalOutputTokens":2349345,"total_cost_usd":1303.46,"total_input_tokens":1367870276,"total_output_tokens":2349345,"total_runtime_s":238983.29079494995,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.714285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b68b14cc721c94b1a086","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2c8bcd72013b959dd6e2d900","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.14144736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":567,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[283];split=full/overall;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":567,"metadata":{"avgScore":0.2927265620525372,"avg_score":0.2927265620525372,"avg_score_percent":29.27265620525372,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2968.62,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":3369212141,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3653426,"passRate":0.14144736842105263,"pass_rate":0.14144736842105263,"pass_rate_percent":14.144736842105262,"passes":22,"runs":149,"runtime":529548.719956005,"source_flags":[],"source_harness":"openclaw","source_row_index":283,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":141,"tokens":null,"totalCostUsd":2968.62,"totalDurationS":529548.719956005,"totalInputTokens":3369212141,"totalOutputTokens":3653426,"total_cost_usd":2968.62,"total_input_tokens":3369212141,"total_output_tokens":3653426,"total_runtime_s":529548.719956005,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.144736842105262,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c3b80427da3559b0a38c","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2418b220e0720060d4bbcc60","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":687,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[343];split=licensed;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":687,"metadata":{"avgScore":0.05454545454545454,"avg_score":0.05454545454545454,"avg_score_percent":5.454545454545454,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":344.43,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":455406090,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":203960,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":79683.10143035022,"source_flags":[],"source_harness":"openclaw","source_row_index":343,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":344.43,"totalDurationS":79683.10143035022,"totalInputTokens":455406090,"totalOutputTokens":203960,"total_cost_usd":344.43,"total_input_tokens":455406090,"total_output_tokens":203960,"total_runtime_s":79683.10143035022,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5ae0383e081cc703afa4","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1ab3bd4d2caaff3e2e65d47d","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1053,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[526];split=unlicensed/full-spectrum;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1053,"metadata":{"avgScore":0.22743563333333333,"avg_score":0.22743563333333333,"avg_score_percent":22.743563333333334,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1191.3,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1373512411,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":1585963,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":5,"runs":51,"runtime":183302.91810718234,"source_flags":[],"source_harness":"openclaw","source_row_index":526,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":47,"tokens":null,"totalCostUsd":1191.3,"totalDurationS":183302.91810718234,"totalInputTokens":1373512411,"totalOutputTokens":1585963,"total_cost_usd":1191.3,"total_input_tokens":1373512411,"total_output_tokens":1585963,"total_runtime_s":183302.91810718234,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-235460fbcff2adba513a","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-03321c3ec83d7c3521f89247","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3-1-pro-preview","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+adj","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1524822695035461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fbe858aafed2e6cb956404291558ef9c35d8a39","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":931,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[465];split=unlicensed;harness=openclaw;model=gemini-3-1-pro-preview;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":931,"metadata":{"avgScore":0.31130806689351526,"avg_score":0.31130806689351526,"avg_score_percent":31.130806689351527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":2624.2,"costSource":"db+adj","cost_source":"db+adj","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2913806051,"metric_id":"pass_rate","metric_role":"primary","model":"gemini-3-1-pro-preview","output_tokens":3449466,"passRate":0.1524822695035461,"pass_rate":0.1524822695035461,"pass_rate_percent":15.24822695035461,"passes":22,"runs":142,"runtime":449865.6185256547,"source_flags":[],"source_harness":"openclaw","source_row_index":465,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":2624.2,"totalDurationS":449865.6185256547,"totalInputTokens":2913806051,"totalOutputTokens":3449466,"total_cost_usd":2624.2,"total_input_tokens":2913806051,"total_output_tokens":3449466,"total_runtime_s":449865.6185256547,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.24822695035461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-63b8064513d13a8156e1","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-eb80142978847acd89ca6619","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2813736456137783,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":100,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[49];split=full;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":100,"metadata":{"avgScore":0.2813736456137783,"avg_score":0.2813736456137783,"avg_score_percent":28.13736456137783,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.78,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":755773178.3333334,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":5896205,"passRate":0.11513157894736842,"pass_rate":0.11513157894736842,"pass_rate_percent":11.513157894736842,"passes":35,"runs":260,"runtime":607138.4432532396,"source_flags":[],"source_harness":"openclaw","source_row_index":49,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":140,"tokens":null,"totalCostUsd":439.78,"totalDurationS":607138.4432532396,"totalInputTokens":755773178.3333334,"totalOutputTokens":5896205,"total_cost_usd":439.78,"total_input_tokens":755773178.3333334,"total_output_tokens":5896205,"total_runtime_s":607138.4432532396,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.13736456137783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e9d4f6e20ccd08afedbe","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8a9a0cc13f64a7150fd86c25","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.061951272828356406,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1166,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[582];split=unlicensed/last-exam;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1166,"metadata":{"avgScore":0.061951272828356406,"avg_score":0.061951272828356406,"avg_score_percent":6.1951272828356405,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":153.78,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":263318872.5,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":2147740,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":50,"runtime":169850.73263209057,"source_flags":[],"source_harness":"openclaw","source_row_index":582,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":153.78,"totalDurationS":169850.73263209057,"totalInputTokens":263318872.5,"totalOutputTokens":2147740,"total_cost_usd":153.78,"total_input_tokens":263318872.5,"total_output_tokens":2147740,"total_runtime_s":169850.73263209057,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.1951272828356405,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-389451118312b3ef5830","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-855c3efa4f9a0d26d99748aa","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2813736456137783,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":572,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[285];split=full/overall;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":572,"metadata":{"avgScore":0.2813736456137783,"avg_score":0.2813736456137783,"avg_score_percent":28.13736456137783,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.78,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":755773178.3333334,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":5896205,"passRate":0.11513157894736842,"pass_rate":0.11513157894736842,"pass_rate_percent":11.513157894736842,"passes":35,"runs":260,"runtime":607138.4432532396,"source_flags":[],"source_harness":"openclaw","source_row_index":285,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":140,"tokens":null,"totalCostUsd":439.78,"totalDurationS":607138.4432532396,"totalInputTokens":755773178.3333334,"totalOutputTokens":5896205,"total_cost_usd":439.78,"total_input_tokens":755773178.3333334,"total_output_tokens":5896205,"total_runtime_s":607138.4432532396,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.13736456137783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd46a02e34f07a553ce4","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6630b63a2534be4e4822f6dc","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3078167601947899,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":816,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[407];split=linux_only;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":816,"metadata":{"avgScore":0.3078167601947899,"avg_score":0.3078167601947899,"avg_score_percent":30.781676019478986,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":212.77,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":361003727.3333333,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":3257192,"passRate":0.12857142857142856,"pass_rate":0.12857142857142856,"pass_rate_percent":12.857142857142856,"passes":27,"runs":193,"runtime":266197.7125626944,"source_flags":[],"source_harness":"openclaw","source_row_index":407,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":100,"tokens":null,"totalCostUsd":212.77,"totalDurationS":266197.7125626944,"totalInputTokens":361003727.3333333,"totalOutputTokens":3257192,"total_cost_usd":212.77,"total_input_tokens":361003727.3333333,"total_output_tokens":3257192,"total_runtime_s":266197.7125626944,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":30.781676019478986,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7b487d11df6ac6b62207","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6376d9eb394d8d3c67b65e29","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.061845741706597165,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":326,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[162];split=full/last-exam;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":326,"metadata":{"avgScore":0.061845741706597165,"avg_score":0.061845741706597165,"avg_score_percent":6.184574170659716,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":203.68,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":353977626.5,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":2529310.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":54,"runtime":205776.51792727213,"source_flags":[],"source_harness":"openclaw","source_row_index":162,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":203.68,"totalDurationS":205776.51792727213,"totalInputTokens":353977626.5,"totalOutputTokens":2529310.5,"total_cost_usd":203.68,"total_input_tokens":353977626.5,"total_output_tokens":2529310.5,"total_runtime_s":205776.51792727213,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.184574170659716,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a5c0b9a50b861e5886d6","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5ad4eff9674f4ae8a3e9dd7d","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.45560963735192905,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1288,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[643];split=unlicensed/near-term;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1288,"metadata":{"avgScore":0.45560963735192905,"avg_score":0.45560963735192905,"avg_score_percent":45.56096373519291,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":183496122.83333334,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":1648850.5,"passRate":0.20149253731343283,"pass_rate":0.20149253731343283,"pass_rate_percent":20.149253731343283,"passes":27,"runs":130,"runtime":224248.87307640666,"source_flags":[],"source_harness":"openclaw","source_row_index":643,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":107.85,"totalDurationS":224248.87307640666,"totalInputTokens":183496122.83333334,"totalOutputTokens":1648850.5,"total_cost_usd":107.85,"total_input_tokens":183496122.83333334,"total_output_tokens":1648850.5,"total_runtime_s":224248.87307640666,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.56096373519291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9ce4180a61ad8e8dddf5","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-542ed245aae98817af4e5c30","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.45560963735192905,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":452,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[225];split=full/near-term;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":452,"metadata":{"avgScore":0.45560963735192905,"avg_score":0.45560963735192905,"avg_score_percent":45.56096373519291,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":183496122.83333334,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":1648850.5,"passRate":0.20149253731343283,"pass_rate":0.20149253731343283,"pass_rate_percent":20.149253731343283,"passes":27,"runs":130,"runtime":224248.87307640666,"source_flags":[],"source_harness":"openclaw","source_row_index":225,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":107.85,"totalDurationS":224248.87307640666,"totalInputTokens":183496122.83333334,"totalOutputTokens":1648850.5,"total_cost_usd":107.85,"total_input_tokens":183496122.83333334,"total_output_tokens":1648850.5,"total_runtime_s":224248.87307640666,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.56096373519291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a362b49785e48ab62517","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3a29bacf1a04aebb3715cfb8","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21708428055845586,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":220,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[109];split=full/full-spectrum;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":220,"metadata":{"avgScore":0.21708428055845586,"avg_score":0.21708428055845586,"avg_score_percent":21.708428055845587,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":165.44,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":282398618,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":2238257.5,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":10,"runs":90,"runtime":222992.05093784933,"source_flags":[],"source_harness":"openclaw","source_row_index":109,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":49,"tokens":null,"totalCostUsd":165.44,"totalDurationS":222992.05093784933,"totalInputTokens":282398618,"totalOutputTokens":2238257.5,"total_cost_usd":165.44,"total_input_tokens":282398618,"total_output_tokens":2238257.5,"total_runtime_s":222992.05093784933,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.708428055845587,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-469b72519fea97292ad0","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1ecab2421971ed7a768dda57","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22985225925256939,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1052,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[525];split=unlicensed/full-spectrum;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1052,"metadata":{"avgScore":0.22985225925256939,"avg_score":0.22985225925256939,"avg_score_percent":22.985225925256938,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":100.65,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":165805606,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":1715714.5,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":10,"runs":82,"runtime":161077.19640037167,"source_flags":[],"source_harness":"openclaw","source_row_index":525,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":44,"tokens":null,"totalCostUsd":100.65,"totalDurationS":161077.19640037167,"totalInputTokens":165805606,"totalOutputTokens":1715714.5,"total_cost_usd":100.65,"total_input_tokens":165805606,"total_output_tokens":1715714.5,"total_runtime_s":161077.19640037167,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.985225925256938,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-96037ce2b369c8bced3b","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1a7a129d83367aa4981c1bca","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2966678841504093,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":938,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[468];split=unlicensed;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":938,"metadata":{"avgScore":0.2966678841504093,"avg_score":0.2966678841504093,"avg_score_percent":29.66678841504093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":348.97,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":592066637.8333334,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":5130397,"passRate":0.12411347517730496,"pass_rate":0.12411347517730496,"pass_rate_percent":12.411347517730496,"passes":35,"runs":251,"runtime":527433.7247366491,"source_flags":[],"source_harness":"openclaw","source_row_index":468,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":348.97,"totalDurationS":527433.7247366491,"totalInputTokens":592066637.8333334,"totalOutputTokens":5130397,"total_cost_usd":348.97,"total_input_tokens":592066637.8333334,"total_output_tokens":5130397,"total_runtime_s":527433.7247366491,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.66678841504093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8d74d0674a1ec0eda36c","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-18de558107c6dd9019ceb285","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2966678841504093,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1410,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[704];split=unlicensed/overall;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1410,"metadata":{"avgScore":0.2966678841504093,"avg_score":0.2966678841504093,"avg_score_percent":29.66678841504093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":348.97,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":592066637.8333334,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":5130397,"passRate":0.12411347517730496,"pass_rate":0.12411347517730496,"pass_rate_percent":12.411347517730496,"passes":35,"runs":251,"runtime":527433.7247366491,"source_flags":[],"source_harness":"openclaw","source_row_index":704,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":348.97,"totalDurationS":527433.7247366491,"totalInputTokens":592066637.8333334,"totalOutputTokens":5130397,"total_cost_usd":348.97,"total_input_tokens":592066637.8333334,"total_output_tokens":5130397,"total_runtime_s":527433.7247366491,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.66678841504093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-da2b83861c54ad078f72","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0231b03d2be25442af8313c9","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08532931528060012,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4fb6f08bab14ee6fd2af782312168e5cc5a2f41","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":668,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[333];split=licensed;harness=openclaw;model=glm-5-1;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":668,"metadata":{"avgScore":0.08532931528060012,"avg_score":0.08532931528060012,"avg_score_percent":8.532931528060011,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":90.81,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":163706540.5,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-1","output_tokens":765808,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":9,"runtime":79704.71851659058,"source_flags":[],"source_harness":"openclaw","source_row_index":333,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":90.81,"totalDurationS":79704.71851659058,"totalInputTokens":163706540.5,"totalOutputTokens":765808,"total_cost_usd":90.81,"total_input_tokens":163706540.5,"total_output_tokens":765808,"total_runtime_s":79704.71851659058,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.532931528060011,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0f8250628321910bbc98","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e45a3ccbc8a7c686fdee5085","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11513157894736842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":571,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[285];split=full/overall;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":571,"metadata":{"avgScore":0.2813736456137783,"avg_score":0.2813736456137783,"avg_score_percent":28.13736456137783,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.78,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":755773178.3333334,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":5896205,"passRate":0.11513157894736842,"pass_rate":0.11513157894736842,"pass_rate_percent":11.513157894736842,"passes":35,"runs":260,"runtime":607138.4432532396,"source_flags":[],"source_harness":"openclaw","source_row_index":285,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":140,"tokens":null,"totalCostUsd":439.78,"totalDurationS":607138.4432532396,"totalInputTokens":755773178.3333334,"totalOutputTokens":5896205,"total_cost_usd":439.78,"total_input_tokens":755773178.3333334,"total_output_tokens":5896205,"total_runtime_s":607138.4432532396,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.513157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-21aef779d7f741eef76a","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b2c66f7a2b874e4bcfcc3f98","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":219,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[109];split=full/full-spectrum;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":219,"metadata":{"avgScore":0.21708428055845586,"avg_score":0.21708428055845586,"avg_score_percent":21.708428055845587,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":165.44,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":282398618,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":2238257.5,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":10,"runs":90,"runtime":222992.05093784933,"source_flags":[],"source_harness":"openclaw","source_row_index":109,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":49,"tokens":null,"totalCostUsd":165.44,"totalDurationS":222992.05093784933,"totalInputTokens":282398618,"totalOutputTokens":2238257.5,"total_cost_usd":165.44,"total_input_tokens":282398618,"total_output_tokens":2238257.5,"total_runtime_s":222992.05093784933,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e16821b8f06ef27a44a2","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9e75c220c8bc6c4dc860bd40","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12411347517730496,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1409,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[704];split=unlicensed/overall;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1409,"metadata":{"avgScore":0.2966678841504093,"avg_score":0.2966678841504093,"avg_score_percent":29.66678841504093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":348.97,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":592066637.8333334,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":5130397,"passRate":0.12411347517730496,"pass_rate":0.12411347517730496,"pass_rate_percent":12.411347517730496,"passes":35,"runs":251,"runtime":527433.7247366491,"source_flags":[],"source_harness":"openclaw","source_row_index":704,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":348.97,"totalDurationS":527433.7247366491,"totalInputTokens":592066637.8333334,"totalOutputTokens":5130397,"total_cost_usd":348.97,"total_input_tokens":592066637.8333334,"total_output_tokens":5130397,"total_runtime_s":527433.7247366491,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.411347517730496,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-79efc8c9616362b732ec","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9e52055c6fdb0257301d258e","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":667,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[333];split=licensed;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":667,"metadata":{"avgScore":0.08532931528060012,"avg_score":0.08532931528060012,"avg_score_percent":8.532931528060011,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":90.81,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":163706540.5,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":765808,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":9,"runtime":79704.71851659058,"source_flags":[],"source_harness":"openclaw","source_row_index":333,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":90.81,"totalDurationS":79704.71851659058,"totalInputTokens":163706540.5,"totalOutputTokens":765808,"total_cost_usd":90.81,"total_input_tokens":163706540.5,"total_output_tokens":765808,"total_runtime_s":79704.71851659058,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7e02dad73b95ab1df9aa","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8a50d3dd085cfcae7f241250","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.20149253731343283,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1287,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[643];split=unlicensed/near-term;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1287,"metadata":{"avgScore":0.45560963735192905,"avg_score":0.45560963735192905,"avg_score_percent":45.56096373519291,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":183496122.83333334,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":1648850.5,"passRate":0.20149253731343283,"pass_rate":0.20149253731343283,"pass_rate_percent":20.149253731343283,"passes":27,"runs":130,"runtime":224248.87307640666,"source_flags":[],"source_harness":"openclaw","source_row_index":643,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":107.85,"totalDurationS":224248.87307640666,"totalInputTokens":183496122.83333334,"totalOutputTokens":1648850.5,"total_cost_usd":107.85,"total_input_tokens":183496122.83333334,"total_output_tokens":1648850.5,"total_runtime_s":224248.87307640666,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.149253731343283,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9b287cd42e8cee00dfc3","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-78ee6b147c0cf0c1b57bd9f9","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11513157894736842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":99,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[49];split=full;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":99,"metadata":{"avgScore":0.2813736456137783,"avg_score":0.2813736456137783,"avg_score_percent":28.13736456137783,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.78,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":755773178.3333334,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":5896205,"passRate":0.11513157894736842,"pass_rate":0.11513157894736842,"pass_rate_percent":11.513157894736842,"passes":35,"runs":260,"runtime":607138.4432532396,"source_flags":[],"source_harness":"openclaw","source_row_index":49,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":140,"tokens":null,"totalCostUsd":439.78,"totalDurationS":607138.4432532396,"totalInputTokens":755773178.3333334,"totalOutputTokens":5896205,"total_cost_usd":439.78,"total_input_tokens":755773178.3333334,"total_output_tokens":5896205,"total_runtime_s":607138.4432532396,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.513157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bd4df0032837a734232b","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-71c308349ae50235503a7549","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":325,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[162];split=full/last-exam;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":325,"metadata":{"avgScore":0.061845741706597165,"avg_score":0.061845741706597165,"avg_score_percent":6.184574170659716,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":203.68,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":353977626.5,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":2529310.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":54,"runtime":205776.51792727213,"source_flags":[],"source_harness":"openclaw","source_row_index":162,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":203.68,"totalDurationS":205776.51792727213,"totalInputTokens":353977626.5,"totalOutputTokens":2529310.5,"total_cost_usd":203.68,"total_input_tokens":353977626.5,"total_output_tokens":2529310.5,"total_runtime_s":205776.51792727213,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-14fec7e5e3be4a74938b","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5a623c1f2386d27d7e68b8bb","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1165,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[582];split=unlicensed/last-exam;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1165,"metadata":{"avgScore":0.061951272828356406,"avg_score":0.061951272828356406,"avg_score_percent":6.1951272828356405,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":153.78,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":263318872.5,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":2147740,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":50,"runtime":169850.73263209057,"source_flags":[],"source_harness":"openclaw","source_row_index":582,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":153.78,"totalDurationS":169850.73263209057,"totalInputTokens":263318872.5,"totalOutputTokens":2147740,"total_cost_usd":153.78,"total_input_tokens":263318872.5,"total_output_tokens":2147740,"total_runtime_s":169850.73263209057,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4da2c3eb0edcac73a388","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-597cd37f27e539a9b6b151cc","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12857142857142856,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":815,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[407];split=linux_only;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":815,"metadata":{"avgScore":0.3078167601947899,"avg_score":0.3078167601947899,"avg_score_percent":30.781676019478986,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":212.77,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":361003727.3333333,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":3257192,"passRate":0.12857142857142856,"pass_rate":0.12857142857142856,"pass_rate_percent":12.857142857142856,"passes":27,"runs":193,"runtime":266197.7125626944,"source_flags":[],"source_harness":"openclaw","source_row_index":407,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":100,"tokens":null,"totalCostUsd":212.77,"totalDurationS":266197.7125626944,"totalInputTokens":361003727.3333333,"totalOutputTokens":3257192,"total_cost_usd":212.77,"total_input_tokens":361003727.3333333,"total_output_tokens":3257192,"total_runtime_s":266197.7125626944,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.857142857142856,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8b9f4a77606fe87ffbd8","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4dc778f7336067df0e81528b","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.20149253731343283,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":451,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[225];split=full/near-term;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":451,"metadata":{"avgScore":0.45560963735192905,"avg_score":0.45560963735192905,"avg_score_percent":45.56096373519291,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":183496122.83333334,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":1648850.5,"passRate":0.20149253731343283,"pass_rate":0.20149253731343283,"pass_rate_percent":20.149253731343283,"passes":27,"runs":130,"runtime":224248.87307640666,"source_flags":[],"source_harness":"openclaw","source_row_index":225,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":107.85,"totalDurationS":224248.87307640666,"totalInputTokens":183496122.83333334,"totalOutputTokens":1648850.5,"total_cost_usd":107.85,"total_input_tokens":183496122.83333334,"total_output_tokens":1648850.5,"total_runtime_s":224248.87307640666,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.149253731343283,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-73351b1414b67ff8d86d","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4b23f2a467dcd1dc6808e6c8","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12411347517730496,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":937,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[468];split=unlicensed;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":937,"metadata":{"avgScore":0.2966678841504093,"avg_score":0.2966678841504093,"avg_score_percent":29.66678841504093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":348.97,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":592066637.8333334,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":5130397,"passRate":0.12411347517730496,"pass_rate":0.12411347517730496,"pass_rate_percent":12.411347517730496,"passes":35,"runs":251,"runtime":527433.7247366491,"source_flags":[],"source_harness":"openclaw","source_row_index":468,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":134,"tokens":null,"totalCostUsd":348.97,"totalDurationS":527433.7247366491,"totalInputTokens":592066637.8333334,"totalOutputTokens":5130397,"total_cost_usd":348.97,"total_input_tokens":592066637.8333334,"total_output_tokens":5130397,"total_runtime_s":527433.7247366491,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.411347517730496,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ed5ef40a48d58483ff7a","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1017cde10f66caf84aa23eb2","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.1","modelRef":"glm-5-1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8551cb9d9cba263db98a45e495f97bf97371a1a9","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1051,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[525];split=unlicensed/full-spectrum;harness=openclaw;model=glm-5-1;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1051,"metadata":{"avgScore":0.22985225925256939,"avg_score":0.22985225925256939,"avg_score_percent":22.985225925256938,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":100.65,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":165805606,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-1","output_tokens":1715714.5,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":10,"runs":82,"runtime":161077.19640037167,"source_flags":[],"source_harness":"openclaw","source_row_index":525,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":44,"tokens":null,"totalCostUsd":100.65,"totalDurationS":161077.19640037167,"totalInputTokens":165805606,"totalOutputTokens":1715714.5,"total_cost_usd":100.65,"total_input_tokens":165805606,"total_output_tokens":1715714.5,"total_runtime_s":161077.19640037167,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-19174fe264c140631ad0","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f5a7998db59d0a7f53713cf7","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4415980756094993,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":758,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[378];split=linux_only;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":758,"metadata":{"avgScore":0.4415980756094993,"avg_score":0.4415980756094993,"avg_score_percent":44.15980756094993,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":980.79,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1190914196,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":6337297,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":25,"runs":104,"runtime":231492,"source_flags":[],"source_harness":"claude_code","source_row_index":378,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":980.79,"totalDurationS":231492,"totalInputTokens":1190914196,"totalOutputTokens":6337297,"total_cost_usd":980.79,"total_input_tokens":1190914196,"total_output_tokens":6337297,"total_runtime_s":231492,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.15980756094993,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-330a21c3279b31ac8135","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c13ab23dea3c83a21eb55f9d","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.41085406808572505,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":530,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[264];split=full/overall;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":530,"metadata":{"avgScore":0.41085406808572505,"avg_score":0.41085406808572505,"avg_score_percent":41.08540680857251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1085.62,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1291344387,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":7729473,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":146,"runtime":387747,"source_flags":[],"source_harness":"claude_code","source_row_index":264,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":1085.62,"totalDurationS":387747,"totalInputTokens":1291344387,"totalOutputTokens":7729473,"total_cost_usd":1085.62,"total_input_tokens":1291344387,"total_output_tokens":7729473,"total_runtime_s":387747,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.08540680857251,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7710e380c6e2abda1499","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a6f5954ef4fdf8ddf3fd8b35","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4293016937371446,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":894,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[446];split=unlicensed;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":894,"metadata":{"avgScore":0.4293016937371446,"avg_score":0.4293016937371446,"avg_score_percent":42.93016937371446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1057.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1263713696,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":7358995,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":136,"runtime":346433,"source_flags":[],"source_harness":"claude_code","source_row_index":446,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1057.31,"totalDurationS":346433,"totalInputTokens":1263713696,"totalOutputTokens":7358995,"total_cost_usd":1057.31,"total_input_tokens":1263713696,"total_output_tokens":7358995,"total_runtime_s":346433,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.93016937371446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-233e84e845a388e8dc9d","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-86e511628c668f98956422df","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.12965036002828909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1108,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[553];split=unlicensed/last-exam;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1108,"metadata":{"avgScore":0.12965036002828909,"avg_score":0.12965036002828909,"avg_score_percent":12.965036002828908,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":617.92,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":763499702,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":2426146,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":27,"runtime":150498,"source_flags":[],"source_harness":"claude_code","source_row_index":553,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":27,"tokens":null,"totalCostUsd":617.92,"totalDurationS":150498,"totalInputTokens":763499702,"totalOutputTokens":2426146,"total_cost_usd":617.92,"total_input_tokens":763499702,"total_output_tokens":2426146,"total_runtime_s":150498,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.965036002828908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1aa296f7bb47737a0758","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-86b7297ba52b15aefd0fdeb6","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.17438904837207417,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":644,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[321];split=licensed;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":644,"metadata":{"avgScore":0.17438904837207417,"avg_score":0.17438904837207417,"avg_score_percent":17.438904837207417,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":28.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":27630691,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":370478,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":41314,"source_flags":[],"source_harness":"claude_code","source_row_index":321,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":28.31,"totalDurationS":41314,"totalInputTokens":27630691,"totalOutputTokens":370478,"total_cost_usd":28.31,"total_input_tokens":27630691,"total_output_tokens":370478,"total_runtime_s":41314,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.438904837207417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-75a3e0b9bf6fe90bf4b2","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-81eb6a50abbf2c2556281ac5","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.41085406808572505,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":58,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[28];split=full;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":58,"metadata":{"avgScore":0.41085406808572505,"avg_score":0.41085406808572505,"avg_score_percent":41.08540680857251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1085.62,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1291344387,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":7729473,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":146,"runtime":387747,"source_flags":[],"source_harness":"claude_code","source_row_index":28,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":1085.62,"totalDurationS":387747,"totalInputTokens":1291344387,"totalOutputTokens":7729473,"total_cost_usd":1085.62,"total_input_tokens":1291344387,"total_output_tokens":7729473,"total_runtime_s":387747,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.08540680857251,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6030d841eaeb6a1580c8","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-76310f4b0d00775d5a61f56c","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.36974012794692096,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":172,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[85];split=full/full-spectrum;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":172,"metadata":{"avgScore":0.36974012794692096,"avg_score":0.36974012794692096,"avg_score_percent":36.974012794692094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":389.75,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":338828819,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":2676225,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":54,"runtime":128703,"source_flags":[],"source_harness":"claude_code","source_row_index":85,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":389.75,"totalDurationS":128703,"totalInputTokens":338828819,"totalOutputTokens":2676225,"total_cost_usd":389.75,"total_input_tokens":338828819,"total_output_tokens":2676225,"total_runtime_s":128703,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.974012794692094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8721de03232b5a5ce342","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7628c82ab13fdc06af896c59","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.12862606139319707,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":272,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[135];split=full/last-exam;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":272,"metadata":{"avgScore":0.12862606139319707,"avg_score":0.12862606139319707,"avg_score_percent":12.862606139319707,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":638.49,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":782672511,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":2708597,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":34,"runtime":172545,"source_flags":[],"source_harness":"claude_code","source_row_index":135,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":34,"tokens":null,"totalCostUsd":638.49,"totalDurationS":172545,"totalInputTokens":782672511,"totalOutputTokens":2708597,"total_cost_usd":638.49,"total_input_tokens":782672511,"total_output_tokens":2708597,"total_runtime_s":172545,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.862606139319707,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd03d0a91eba0aa8eef2","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-4d825d81b1f66cd7f08cdfec","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.6033126287817978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1258,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[628];split=unlicensed/near-term;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1258,"metadata":{"avgScore":0.6033126287817978,"avg_score":0.6033126287817978,"avg_score_percent":60.33126287817978,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":204.15,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":187022399,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":2708776,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":66,"runtime":117886,"source_flags":[],"source_harness":"claude_code","source_row_index":628,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":204.15,"totalDurationS":117886,"totalInputTokens":187022399,"totalOutputTokens":2708776,"total_cost_usd":204.15,"total_input_tokens":187022399,"total_output_tokens":2708776,"total_runtime_s":117886,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.33126287817978,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bd1843fb4e0a5cfd3f85","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2861afe20e464a4fa94f26cf","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.36831414074161306,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1010,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[504];split=unlicensed/full-spectrum;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1010,"metadata":{"avgScore":0.36831414074161306,"avg_score":0.36831414074161306,"avg_score_percent":36.83141407416131,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":375.15,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":323665898,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":2488434,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":49,"runtime":105972,"source_flags":[],"source_harness":"claude_code","source_row_index":504,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":375.15,"totalDurationS":105972,"totalInputTokens":323665898,"totalOutputTokens":2488434,"total_cost_usd":375.15,"total_input_tokens":323665898,"total_output_tokens":2488434,"total_runtime_s":105972,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.83141407416131,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b3fa32d598b6c1bc8ff7","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-219b3cc71bcda401d5cf88f0","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4293016937371446,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1366,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[682];split=unlicensed/overall;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1366,"metadata":{"avgScore":0.4293016937371446,"avg_score":0.4293016937371446,"avg_score_percent":42.93016937371446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1057.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1263713696,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":7358995,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":136,"runtime":346433,"source_flags":[],"source_harness":"claude_code","source_row_index":682,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1057.31,"totalDurationS":346433,"totalInputTokens":1263713696,"totalOutputTokens":7358995,"total_cost_usd":1057.31,"total_input_tokens":1263713696,"total_output_tokens":7358995,"total_runtime_s":346433,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.93016937371446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b859c8017dbebe20af3e","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1bd2b2ed2032cd8b58ea6274","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.6033126287817978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b637d843fa03a483ac4961d5629d55f1e9a23e7d","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":422,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[210];split=full/near-term;harness=claude_code;model=glm-5-2;variant=max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":422,"metadata":{"avgScore":0.6033126287817978,"avg_score":0.6033126287817978,"avg_score_percent":60.33126287817978,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":204.15,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":187022399,"metric_id":"avg_score","metric_role":"secondary","model":"glm-5-2","output_tokens":2708776,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":66,"runtime":117886,"source_flags":[],"source_harness":"claude_code","source_row_index":210,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":204.15,"totalDurationS":117886,"totalInputTokens":187022399,"totalOutputTokens":2708776,"total_cost_usd":204.15,"total_input_tokens":187022399,"total_output_tokens":2708776,"total_runtime_s":117886,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.33126287817978,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5f68c1d2b6f991f94d16","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-fba8a3365033e223c3eefdc8","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":643,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[321];split=licensed;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":643,"metadata":{"avgScore":0.17438904837207417,"avg_score":0.17438904837207417,"avg_score_percent":17.438904837207417,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":28.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":27630691,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":370478,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":41314,"source_flags":[],"source_harness":"claude_code","source_row_index":321,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":28.31,"totalDurationS":41314,"totalInputTokens":27630691,"totalOutputTokens":370478,"total_cost_usd":28.31,"total_input_tokens":27630691,"total_output_tokens":370478,"total_runtime_s":41314,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a381481a6b131155ab76","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ece8ac165ff96318da5cb5df","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2198581560283688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1365,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[682];split=unlicensed/overall;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1365,"metadata":{"avgScore":0.4293016937371446,"avg_score":0.4293016937371446,"avg_score_percent":42.93016937371446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1057.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1263713696,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":7358995,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":136,"runtime":346433,"source_flags":[],"source_harness":"claude_code","source_row_index":682,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1057.31,"totalDurationS":346433,"totalInputTokens":1263713696,"totalOutputTokens":7358995,"total_cost_usd":1057.31,"total_input_tokens":1263713696,"total_output_tokens":7358995,"total_runtime_s":346433,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.98581560283688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8a72b4cd9a412e038162","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ceb45305b7e4ad3cc460ea19","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.20394736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":57,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[28];split=full;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":57,"metadata":{"avgScore":0.41085406808572505,"avg_score":0.41085406808572505,"avg_score_percent":41.08540680857251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1085.62,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1291344387,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":7729473,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":146,"runtime":387747,"source_flags":[],"source_harness":"claude_code","source_row_index":28,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":1085.62,"totalDurationS":387747,"totalInputTokens":1291344387,"totalOutputTokens":7729473,"total_cost_usd":1085.62,"total_input_tokens":1291344387,"total_output_tokens":7729473,"total_runtime_s":387747,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.394736842105264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8d6799d525c8fda9891c","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b51a88d957082b0f6af8ef4a","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3283582089552239,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1257,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[628];split=unlicensed/near-term;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1257,"metadata":{"avgScore":0.6033126287817978,"avg_score":0.6033126287817978,"avg_score_percent":60.33126287817978,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":204.15,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":187022399,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":2708776,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":66,"runtime":117886,"source_flags":[],"source_harness":"claude_code","source_row_index":628,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":204.15,"totalDurationS":117886,"totalInputTokens":187022399,"totalOutputTokens":2708776,"total_cost_usd":204.15,"total_input_tokens":187022399,"total_output_tokens":2708776,"total_runtime_s":117886,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.83582089552239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ed8bf03ed421de37db5f","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b0fbbd82d997aef2389f2729","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3283582089552239,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":421,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[210];split=full/near-term;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":421,"metadata":{"avgScore":0.6033126287817978,"avg_score":0.6033126287817978,"avg_score_percent":60.33126287817978,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":204.15,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":187022399,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":2708776,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":66,"runtime":117886,"source_flags":[],"source_harness":"claude_code","source_row_index":210,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":204.15,"totalDurationS":117886,"totalInputTokens":187022399,"totalOutputTokens":2708776,"total_cost_usd":204.15,"total_input_tokens":187022399,"total_output_tokens":2708776,"total_runtime_s":117886,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.83582089552239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-32983c9f9f5e30210fca","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-88caa6a764cc9aa82fd01d18","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1009,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[504];split=unlicensed/full-spectrum;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1009,"metadata":{"avgScore":0.36831414074161306,"avg_score":0.36831414074161306,"avg_score_percent":36.83141407416131,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":375.15,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":323665898,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":2488434,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":49,"runtime":105972,"source_flags":[],"source_harness":"claude_code","source_row_index":504,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":375.15,"totalDurationS":105972,"totalInputTokens":323665898,"totalOutputTokens":2488434,"total_cost_usd":375.15,"total_input_tokens":323665898,"total_output_tokens":2488434,"total_runtime_s":105972,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2880631063f01cc881d6","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-84e27dc61096ded1a59adab5","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":271,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[135];split=full/last-exam;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":271,"metadata":{"avgScore":0.12862606139319707,"avg_score":0.12862606139319707,"avg_score_percent":12.862606139319707,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":638.49,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":782672511,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":2708597,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":34,"runtime":172545,"source_flags":[],"source_harness":"claude_code","source_row_index":135,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":34,"tokens":null,"totalCostUsd":638.49,"totalDurationS":172545,"totalInputTokens":782672511,"totalOutputTokens":2708597,"total_cost_usd":638.49,"total_input_tokens":782672511,"total_output_tokens":2708597,"total_runtime_s":172545,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-90447bab3dfc33e14dcc","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3d3cc9280d742d4e8b0df11a","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1107,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[553];split=unlicensed/last-exam;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1107,"metadata":{"avgScore":0.12965036002828909,"avg_score":0.12965036002828909,"avg_score_percent":12.965036002828908,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":617.92,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":763499702,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":2426146,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":27,"runtime":150498,"source_flags":[],"source_harness":"claude_code","source_row_index":553,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":27,"tokens":null,"totalCostUsd":617.92,"totalDurationS":150498,"totalInputTokens":763499702,"totalOutputTokens":2426146,"total_cost_usd":617.92,"total_input_tokens":763499702,"total_output_tokens":2426146,"total_runtime_s":150498,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bc9539c9fe1be38369d6","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3caa7d09414aaa949c067317","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.20394736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":529,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[264];split=full/overall;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":529,"metadata":{"avgScore":0.41085406808572505,"avg_score":0.41085406808572505,"avg_score_percent":41.08540680857251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1085.62,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1291344387,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":7729473,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":146,"runtime":387747,"source_flags":[],"source_harness":"claude_code","source_row_index":264,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":1085.62,"totalDurationS":387747,"totalInputTokens":1291344387,"totalOutputTokens":7729473,"total_cost_usd":1085.62,"total_input_tokens":1291344387,"total_output_tokens":7729473,"total_runtime_s":387747,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.394736842105264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-df84b78fd8830154b2cd","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-32f5513bc1dd696e0d93d3e7","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.23809523809523808,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":757,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[378];split=linux_only;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":757,"metadata":{"avgScore":0.4415980756094993,"avg_score":0.4415980756094993,"avg_score_percent":44.15980756094993,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":980.79,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1190914196,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":6337297,"passRate":0.23809523809523808,"pass_rate":0.23809523809523808,"pass_rate_percent":23.809523809523807,"passes":25,"runs":104,"runtime":231492,"source_flags":[],"source_harness":"claude_code","source_row_index":378,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":980.79,"totalDurationS":231492,"totalInputTokens":1190914196,"totalOutputTokens":6337297,"total_cost_usd":980.79,"total_input_tokens":1190914196,"total_output_tokens":6337297,"total_runtime_s":231492,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.809523809523807,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-371d7764922946ea903d","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-23e8b1d0b66f03bcef36527b","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2198581560283688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":893,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[446];split=unlicensed;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":893,"metadata":{"avgScore":0.4293016937371446,"avg_score":0.4293016937371446,"avg_score_percent":42.93016937371446,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1057.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":1263713696,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":7358995,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":136,"runtime":346433,"source_flags":[],"source_harness":"claude_code","source_row_index":446,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":136,"tokens":null,"totalCostUsd":1057.31,"totalDurationS":346433,"totalInputTokens":1263713696,"totalOutputTokens":7358995,"total_cost_usd":1057.31,"total_input_tokens":1263713696,"total_output_tokens":7358995,"total_runtime_s":346433,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.98581560283688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d9dd84364bf1b23ee8c5","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-183c430d607d06af0fb0d633","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GLM-5.2","modelRef":"glm-5-2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9a6bdfd0c41e34cb1f94ff55acac3977006d6de","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":171,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[85];split=full/full-spectrum;harness=claude_code;model=glm-5-2;variant=max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":171,"metadata":{"avgScore":0.36974012794692096,"avg_score":0.36974012794692096,"avg_score_percent":36.974012794692094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":389.75,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"max","input_tokens":338828819,"metric_id":"pass_rate","metric_role":"primary","model":"glm-5-2","output_tokens":2676225,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":54,"runtime":128703,"source_flags":[],"source_harness":"claude_code","source_row_index":85,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":389.75,"totalDurationS":128703,"totalInputTokens":338828819,"totalOutputTokens":2676225,"total_cost_usd":389.75,"total_input_tokens":338828819,"total_output_tokens":2676225,"total_runtime_s":128703,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d3cd2b1d8cb9f06e7c80","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f75f854e88cb0b68ed42eadb","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.133367903588527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":588,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[293];split=full/overall;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":588,"metadata":{"avgScore":0.133367903588527,"avg_score":0.133367903588527,"avg_score_percent":13.3367903588527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":134.93,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":206835017.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":3239534.5,"passRate":0.07236842105263158,"pass_rate":0.07236842105263158,"pass_rate_percent":7.236842105263158,"passes":11,"runs":144,"runtime":147936.49156970065,"source_flags":[],"source_harness":"codex","source_row_index":293,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":134.93,"totalDurationS":147936.49156970065,"totalInputTokens":206835017.5,"totalOutputTokens":3239534.5,"total_cost_usd":134.93,"total_input_tokens":206835017.5,"total_output_tokens":3239534.5,"total_runtime_s":147936.49156970065,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.3367903588527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-142ba59eb4b3fc638652","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e99a1a50e53d55d4187bc1cc","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08134244963505957,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":314,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[156];split=full/last-exam;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":314,"metadata":{"avgScore":0.08134244963505957,"avg_score":0.08134244963505957,"avg_score_percent":8.134244963505957,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":100.25,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":321376028,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":540717,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":63550.53847987812,"source_flags":[],"source_harness":"ale_claw","source_row_index":156,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":100.25,"totalDurationS":63550.53847987812,"totalInputTokens":321376028,"totalOutputTokens":540717,"total_cost_usd":100.25,"total_input_tokens":321376028,"total_output_tokens":540717,"total_runtime_s":63550.53847987812,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.134244963505957,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-16e43bc47fe735353437","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e79e204361e9113a5fdc4181","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4184820514506088,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":778,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[388];split=linux_only;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":778,"metadata":{"avgScore":0.4184820514506088,"avg_score":0.4184820514506088,"avg_score_percent":41.84820514506088,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":169.13,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":285789922,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":4866454.666666667,"passRate":0.2253968253968254,"pass_rate":0.2253968253968254,"pass_rate_percent":22.53968253968254,"passes":32,"runs":130,"runtime":180443.07324218247,"source_flags":[],"source_harness":"openclaw","source_row_index":388,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":169.13,"totalDurationS":180443.07324218247,"totalInputTokens":285789922,"totalOutputTokens":4866454.666666667,"total_cost_usd":169.13,"total_input_tokens":285789922,"total_output_tokens":4866454.666666667,"total_runtime_s":180443.07324218247,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":41.84820514506088,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8e07d1f6468e58b3062","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e21fc38dd28c0835cf203ed3","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3434147041110992,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":182,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[90];split=full/full-spectrum;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":182,"metadata":{"avgScore":0.3434147041110992,"avg_score":0.3434147041110992,"avg_score_percent":34.34147041110992,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":122.6,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":239603010.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2920663.6666666665,"passRate":0.19393939393939394,"pass_rate":0.19393939393939394,"pass_rate_percent":19.393939393939394,"passes":13,"runs":71,"runtime":208103.63729115008,"source_flags":[],"source_harness":"openclaw","source_row_index":90,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":122.6,"totalDurationS":208103.63729115008,"totalInputTokens":239603010.16666666,"totalOutputTokens":2920663.6666666665,"total_cost_usd":122.6,"total_input_tokens":239603010.16666666,"total_output_tokens":2920663.6666666665,"total_runtime_s":208103.63729115008,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":34.34147041110992,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-67363f1b74c362f3f743","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-df3452ecb14ddf8a79300f29","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.23974389265492613,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":470,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[234];split=full/near-term;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":470,"metadata":{"avgScore":0.23974389265492613,"avg_score":0.23974389265492613,"avg_score_percent":23.97438926549261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":56.14,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":85510462,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1248047,"passRate":0.13432835820895522,"pass_rate":0.13432835820895522,"pass_rate_percent":13.432835820895523,"passes":9,"runs":68,"runtime":84571.9589642135,"source_flags":[],"source_harness":"codex","source_row_index":234,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":56.14,"totalDurationS":84571.9589642135,"totalInputTokens":85510462,"totalOutputTokens":1248047,"total_cost_usd":56.14,"total_input_tokens":85510462,"total_output_tokens":1248047,"total_runtime_s":84571.9589642135,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.97438926549261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4cef4f432e9244002873","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-dd9df60aff8e056a8ee797dd","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13116963841633533,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":836,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[417];split=linux_only;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":836,"metadata":{"avgScore":0.13116963841633533,"avg_score":0.13116963841633533,"avg_score_percent":13.116963841633533,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":62.4,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":85353349,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1849844,"passRate":0.06666666666666667,"pass_rate":0.06666666666666667,"pass_rate_percent":6.666666666666667,"passes":7,"runs":102,"runtime":39510.33435794874,"source_flags":[],"source_harness":"codex","source_row_index":417,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":62.4,"totalDurationS":39510.33435794874,"totalInputTokens":85353349,"totalOutputTokens":1849844,"total_cost_usd":62.4,"total_input_tokens":85353349,"total_output_tokens":1849844,"total_runtime_s":39510.33435794874,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.116963841633533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6e087fb0707a21c747c2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dba4b47ff595df54c1caa64f","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22512809050649008,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1056,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[527];split=unlicensed/full-spectrum;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1056,"metadata":{"avgScore":0.22512809050649008,"avg_score":0.22512809050649008,"avg_score_percent":22.512809050649008,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":93.03,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":301408364,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":615102,"passRate":0.08,"pass_rate":0.08,"pass_rate_percent":8.0,"passes":4,"runs":49,"runtime":40180.566417631,"source_flags":[],"source_harness":"ale_claw","source_row_index":527,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":93.03,"totalDurationS":40180.566417631,"totalInputTokens":301408364,"totalOutputTokens":615102,"total_cost_usd":93.03,"total_input_tokens":301408364,"total_output_tokens":615102,"total_runtime_s":40180.566417631,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.512809050649008,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-db0cbad1c0d49ad892a9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d9cc1e70f94f35dad0191ed1","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.23974389265492613,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1306,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[652];split=unlicensed/near-term;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1306,"metadata":{"avgScore":0.23974389265492613,"avg_score":0.23974389265492613,"avg_score_percent":23.97438926549261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":56.14,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":85510462,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1248047,"passRate":0.13432835820895522,"pass_rate":0.13432835820895522,"pass_rate_percent":13.432835820895523,"passes":9,"runs":68,"runtime":84571.9589642135,"source_flags":[],"source_harness":"codex","source_row_index":652,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":56.14,"totalDurationS":84571.9589642135,"totalInputTokens":85510462,"totalOutputTokens":1248047,"total_cost_usd":56.14,"total_input_tokens":85510462,"total_output_tokens":1248047,"total_runtime_s":84571.9589642135,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.97438926549261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-950de140175c3b25403e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ca4466f40f4c890bfbca564a","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.45373662346275323,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":454,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[226];split=full/near-term;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":454,"metadata":{"avgScore":0.45373662346275323,"avg_score":0.45373662346275323,"avg_score_percent":45.373662346275324,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":66.17,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":178595579,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":684821,"passRate":0.208955223880597,"pass_rate":0.208955223880597,"pass_rate_percent":20.8955223880597,"passes":14,"runs":66,"runtime":62517.515991543354,"source_flags":[],"source_harness":"ale_claw","source_row_index":226,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":66.17,"totalDurationS":62517.515991543354,"totalInputTokens":178595579,"totalOutputTokens":684821,"total_cost_usd":66.17,"total_input_tokens":178595579,"total_output_tokens":684821,"total_runtime_s":62517.515991543354,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.373662346275324,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6cb51d478545b3d8f2e6","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bbdbfdfc39d685ae856ba6e2","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.33810777616126964,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":802,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[400];split=linux_only;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":802,"metadata":{"avgScore":0.33810777616126964,"avg_score":0.33810777616126964,"avg_score_percent":33.81077761612696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":168.38,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":541326506,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1131556,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":14,"runs":103,"runtime":95462.72998667358,"source_flags":[],"source_harness":"ale_claw","source_row_index":400,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":168.38,"totalDurationS":95462.72998667358,"totalInputTokens":541326506,"totalOutputTokens":1131556,"total_cost_usd":168.38,"total_input_tokens":541326506,"total_output_tokens":1131556,"total_runtime_s":95462.72998667358,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.81077761612696,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f96be77fc1f92b9f84ea","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ad41805c01ce80c6bce61364","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.041695254559251674,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":698,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[348];split=licensed;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":698,"metadata":{"avgScore":0.041695254559251674,"avg_score":0.041695254559251674,"avg_score_percent":4.169525455925167,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":32.38,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":70118711,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":504819,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":82357.47953877086,"source_flags":[],"source_harness":"openclaw","source_row_index":348,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":32.38,"totalDurationS":82357.47953877086,"totalInputTokens":70118711,"totalOutputTokens":504819,"total_cost_usd":32.38,"total_input_tokens":70118711,"total_output_tokens":504819,"total_runtime_s":82357.47953877086,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.169525455925167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b24fc2b9807de665de77","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ab91d540281ec06f82dc7e09","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.578007915138166,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1272,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[635];split=unlicensed/near-term;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1272,"metadata":{"avgScore":0.578007915138166,"avg_score":0.578007915138166,"avg_score_percent":57.8007915138166,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.88,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86038352.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2379425,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":30,"runs":84,"runtime":147335.50355529715,"source_flags":[],"source_harness":"openclaw","source_row_index":635,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":68.88,"totalDurationS":147335.50355529715,"totalInputTokens":86038352.5,"totalOutputTokens":2379425,"total_cost_usd":68.88,"total_input_tokens":86038352.5,"total_output_tokens":2379425,"total_runtime_s":147335.50355529715,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":57.8007915138166,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a9e1ce022cc8c6ac35c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a4613cd1feaa3790ed62f901","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.578007915138166,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":436,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[217];split=full/near-term;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":436,"metadata":{"avgScore":0.578007915138166,"avg_score":0.578007915138166,"avg_score_percent":57.8007915138166,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.88,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86038352.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2379425,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":30,"runs":84,"runtime":147335.50355529715,"source_flags":[],"source_harness":"openclaw","source_row_index":217,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":68.88,"totalDurationS":147335.50355529715,"totalInputTokens":86038352.5,"totalOutputTokens":2379425,"total_cost_usd":68.88,"total_input_tokens":86038352.5,"total_output_tokens":2379425,"total_runtime_s":147335.50355529715,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":57.8007915138166,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e2c02724128ebda4425a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8d8f8fc07a2282e5fb97d237","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13781504500323477,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1424,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[711];split=unlicensed/overall;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1424,"metadata":{"avgScore":0.13781504500323477,"avg_score":0.13781504500323477,"avg_score_percent":13.781504500323477,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":117.05,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":168779753.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2988511.5,"passRate":0.07801418439716312,"pass_rate":0.07801418439716312,"pass_rate_percent":7.801418439716312,"passes":11,"runs":137,"runtime":121711.8690922095,"source_flags":[],"source_harness":"codex","source_row_index":711,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":117.05,"totalDurationS":121711.8690922095,"totalInputTokens":168779753.5,"totalOutputTokens":2988511.5,"total_cost_usd":117.05,"total_input_tokens":168779753.5,"total_output_tokens":2988511.5,"total_runtime_s":121711.8690922095,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.781504500323477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cda80dba646fd4112587","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8d0a7dbd83792c2b18c564f8","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.133367903588527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":116,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[57];split=full;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":116,"metadata":{"avgScore":0.133367903588527,"avg_score":0.133367903588527,"avg_score_percent":13.3367903588527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":134.93,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":206835017.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":3239534.5,"passRate":0.07236842105263158,"pass_rate":0.07236842105263158,"pass_rate_percent":7.236842105263158,"passes":11,"runs":144,"runtime":147936.49156970065,"source_flags":[],"source_harness":"codex","source_row_index":57,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":134.93,"totalDurationS":147936.49156970065,"totalInputTokens":206835017.5,"totalOutputTokens":3239534.5,"total_cost_usd":134.93,"total_input_tokens":206835017.5,"total_output_tokens":3239534.5,"total_runtime_s":147936.49156970065,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.3367903588527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-959303ccc0c24c7457c0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-74d9d10b2239337d2ae9f01e","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3732607831603129,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":78,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[38];split=full;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":78,"metadata":{"avgScore":0.3732607831603129,"avg_score":0.3732607831603129,"avg_score_percent":37.32607831603129,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":274.1,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":488753872.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":7283825.166666667,"passRate":0.2050438596491228,"pass_rate":0.2050438596491228,"pass_rate_percent":20.50438596491228,"passes":40,"runs":179,"runtime":459960.9185098016,"source_flags":[],"source_harness":"openclaw","source_row_index":38,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":274.1,"totalDurationS":459960.9185098016,"totalInputTokens":488753872.1666667,"totalOutputTokens":7283825.166666667,"total_cost_usd":274.1,"total_input_tokens":488753872.1666667,"total_output_tokens":7283825.166666667,"total_runtime_s":459960.9185098016,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.32607831603129,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d166bf79a4070a90783d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6f65a58ce6eab3b6b3829589","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1190,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[594];split=unlicensed/last-exam;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1190,"metadata":{"avgScore":0,"avg_score":0,"avg_score_percent":0.0,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":40.25,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":58038584.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1069865.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":23240.311491120607,"source_flags":[],"source_harness":"codex","source_row_index":594,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":40.25,"totalDurationS":23240.311491120607,"totalInputTokens":58038584.5,"totalOutputTokens":1069865.5,"total_cost_usd":40.25,"total_input_tokens":58038584.5,"total_output_tokens":1069865.5,"total_runtime_s":23240.311491120607,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2d2bb3d36f643fe7aedb","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6b1d72c49587a54b0b80448e","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13781504500323477,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":952,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[475];split=unlicensed;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":952,"metadata":{"avgScore":0.13781504500323477,"avg_score":0.13781504500323477,"avg_score_percent":13.781504500323477,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":117.05,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":168779753.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2988511.5,"passRate":0.07801418439716312,"pass_rate":0.07801418439716312,"pass_rate_percent":7.801418439716312,"passes":11,"runs":137,"runtime":121711.8690922095,"source_flags":[],"source_harness":"codex","source_row_index":475,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":117.05,"totalDurationS":121711.8690922095,"totalInputTokens":168779753.5,"totalOutputTokens":2988511.5,"total_cost_usd":117.05,"total_input_tokens":168779753.5,"total_output_tokens":2988511.5,"total_runtime_s":121711.8690922095,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.781504500323477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8326dd5db2882d68621","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6621c9971a7e5c5bcb06e306","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2863832387982169,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":98,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[48];split=full;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":98,"metadata":{"avgScore":0.2863832387982169,"avg_score":0.2863832387982169,"avg_score_percent":28.63832387982169,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":335.05,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1070761845,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2049232,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":143,"runtime":199675.9292053005,"source_flags":[],"source_harness":"ale_claw","source_row_index":48,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":335.05,"totalDurationS":199675.9292053005,"totalInputTokens":1070761845,"totalOutputTokens":2049232,"total_cost_usd":335.05,"total_input_tokens":1070761845,"total_output_tokens":2049232,"total_runtime_s":199675.9292053005,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.63832387982169,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1fb88a5bc689cf3ab41","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5619a46a1b5ef63137314d8d","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10303376953774213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1130,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[564];split=unlicensed/last-exam;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1130,"metadata":{"avgScore":0.10303376953774213,"avg_score":0.10303376953774213,"avg_score_percent":10.303376953774212,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":91.02,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":296584939,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":472813,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":28,"runtime":56400.74217087712,"source_flags":[],"source_harness":"ale_claw","source_row_index":564,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":91.02,"totalDurationS":56400.74217087712,"totalInputTokens":296584939,"totalOutputTokens":472813,"total_cost_usd":91.02,"total_input_tokens":296584939,"total_output_tokens":472813,"total_runtime_s":56400.74217087712,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.303376953774212,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a649ac81ea526358e5a9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-53998c78b891fc4077654743","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.39912759744833903,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1386,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[692];split=unlicensed/overall;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1386,"metadata":{"avgScore":0.39912759744833903,"avg_score":0.39912759744833903,"avg_score_percent":39.912759744833906,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":241.72,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":418635161.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":6779006.166666667,"passRate":0.22104018912529552,"pass_rate":0.22104018912529552,"pass_rate_percent":22.10401891252955,"passes":40,"runs":172,"runtime":377603.43897103076,"source_flags":[],"source_harness":"openclaw","source_row_index":692,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":241.72,"totalDurationS":377603.43897103076,"totalInputTokens":418635161.1666667,"totalOutputTokens":6779006.166666667,"total_cost_usd":241.72,"total_input_tokens":418635161.1666667,"total_output_tokens":6779006.166666667,"total_runtime_s":377603.43897103076,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":39.912759744833906,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-17c05b76a06afe26cae9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4fd09e5278d8a4f83b6806fd","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.45373662346275323,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1290,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[644];split=unlicensed/near-term;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1290,"metadata":{"avgScore":0.45373662346275323,"avg_score":0.45373662346275323,"avg_score_percent":45.373662346275324,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":66.17,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":178595579,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":684821,"passRate":0.208955223880597,"pass_rate":0.208955223880597,"pass_rate_percent":20.8955223880597,"passes":14,"runs":66,"runtime":62517.515991543354,"source_flags":[],"source_harness":"ale_claw","source_row_index":644,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":66.17,"totalDurationS":62517.515991543354,"totalInputTokens":178595579,"totalOutputTokens":684821,"total_cost_usd":66.17,"total_input_tokens":178595579,"total_output_tokens":684821,"total_runtime_s":62517.515991543354,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.373662346275324,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ccca855843d10100716b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4b3a285c5a0e75fe9b707b5d","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.39912759744833903,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":914,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[456];split=unlicensed;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":914,"metadata":{"avgScore":0.39912759744833903,"avg_score":0.39912759744833903,"avg_score_percent":39.912759744833906,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":241.72,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":418635161.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":6779006.166666667,"passRate":0.22104018912529552,"pass_rate":0.22104018912529552,"pass_rate_percent":22.10401891252955,"passes":40,"runs":172,"runtime":377603.43897103076,"source_flags":[],"source_harness":"openclaw","source_row_index":456,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":241.72,"totalDurationS":377603.43897103076,"totalInputTokens":418635161.1666667,"totalOutputTokens":6779006.166666667,"total_cost_usd":241.72,"total_input_tokens":418635161.1666667,"total_output_tokens":6779006.166666667,"total_runtime_s":377603.43897103076,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":39.912759744833906,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4329fbbdd93441845ca0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-268361a7e21530403db0b820","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03272727272727272,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":700,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[349];split=licensed;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":700,"metadata":{"avgScore":0.03272727272727272,"avg_score":0.03272727272727272,"avg_score_percent":3.2727272727272725,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":87.74,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":299479453,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":328146,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":6,"runtime":43091.204194458005,"source_flags":[],"source_harness":"ale_claw","source_row_index":349,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":87.74,"totalDurationS":43091.204194458005,"totalInputTokens":299479453,"totalOutputTokens":328146,"total_cost_usd":87.74,"total_input_tokens":299479453,"total_output_tokens":328146,"total_runtime_s":43091.204194458005,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.2727272727272725,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f70db4f504f8ddb2a1c3","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2521c0d59e4257929a205015","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.30617200210871603,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":934,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[466];split=unlicensed;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":934,"metadata":{"avgScore":0.30617200210871603,"avg_score":0.30617200210871603,"avg_score_percent":30.617200210871605,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":247.31,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":771282392,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1721086,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":137,"runtime":156584.72501084246,"source_flags":[],"source_harness":"ale_claw","source_row_index":466,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":247.31,"totalDurationS":156584.72501084246,"totalInputTokens":771282392,"totalOutputTokens":1721086,"total_cost_usd":247.31,"total_input_tokens":771282392,"total_output_tokens":1721086,"total_runtime_s":156584.72501084246,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":30.617200210871605,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-73b665905fe15ac879c1","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1cb2250052a0adfe6445a184","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22938917318771823,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":216,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[107];split=full/full-spectrum;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":216,"metadata":{"avgScore":0.22938917318771823,"avg_score":0.22938917318771823,"avg_score_percent":22.93891731877182,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":178.85,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":595700026,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":929990,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":5,"runs":54,"runtime":81361.152385172,"source_flags":[],"source_harness":"ale_claw","source_row_index":107,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":178.85,"totalDurationS":81361.152385172,"totalInputTokens":595700026,"totalOutputTokens":929990,"total_cost_usd":178.85,"total_input_tokens":595700026,"total_output_tokens":929990,"total_runtime_s":81361.152385172,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.93891731877182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cb6347c32b1544b9fd90","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1c19b3cd822fe07c2ecc27f4","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":674,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[336];split=licensed;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":674,"metadata":{"avgScore":0.07636363636363636,"avg_score":0.07636363636363636,"avg_score_percent":7.636363636363636,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":17.87,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":38055264,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":251023,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":26224.622477491153,"source_flags":[],"source_harness":"codex","source_row_index":336,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":17.87,"totalDurationS":26224.622477491153,"totalInputTokens":38055264,"totalOutputTokens":251023,"total_cost_usd":17.87,"total_input_tokens":38055264,"total_output_tokens":251023,"total_runtime_s":26224.622477491153,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.636363636363636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-86cdf56d7512a7cd63cd","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-16feba65573dfdffb50afa18","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.07318078210639797,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":320,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[159];split=full/last-exam;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":320,"metadata":{"avgScore":0.07318078210639797,"avg_score":0.07318078210639797,"avg_score_percent":7.318078210639797,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":106.76,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":206932842.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2518416,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":37,"runtime":130535.2900615723,"source_flags":[],"source_harness":"openclaw","source_row_index":159,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":106.76,"totalDurationS":130535.2900615723,"totalInputTokens":206932842.16666666,"totalOutputTokens":2518416,"total_cost_usd":106.76,"total_input_tokens":206932842.16666666,"total_output_tokens":2518416,"total_runtime_s":130535.2900615723,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":7.318078210639797,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7ec01af3c556ad36f886","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-13e915170f768dd13f09ca9e","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2863832387982169,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":570,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[284];split=full/overall;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":570,"metadata":{"avgScore":0.2863832387982169,"avg_score":0.2863832387982169,"avg_score_percent":28.63832387982169,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":335.05,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1070761845,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2049232,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":143,"runtime":199675.9292053005,"source_flags":[],"source_harness":"ale_claw","source_row_index":284,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":335.05,"totalDurationS":199675.9292053005,"totalInputTokens":1070761845,"totalOutputTokens":2049232,"total_cost_usd":335.05,"total_input_tokens":1070761845,"total_output_tokens":2049232,"total_runtime_s":199675.9292053005,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.63832387982169,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f9ac48a91f19daa0bdcf","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-11f76f47625caf8eb47653d1","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.30617200210871603,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1406,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[702];split=unlicensed/overall;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1406,"metadata":{"avgScore":0.30617200210871603,"avg_score":0.30617200210871603,"avg_score_percent":30.617200210871605,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":247.31,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":771282392,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1721086,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":137,"runtime":156584.72501084246,"source_flags":[],"source_harness":"ale_claw","source_row_index":702,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":247.31,"totalDurationS":156584.72501084246,"totalInputTokens":771282392,"totalOutputTokens":1721086,"total_cost_usd":247.31,"total_input_tokens":771282392,"total_output_tokens":1721086,"total_runtime_s":156584.72501084246,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":30.617200210871605,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e83ef11a153724affc20","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-105eb85459e33883ea1096a4","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09269565733477077,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1142,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[570];split=unlicensed/last-exam;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1142,"metadata":{"avgScore":0.09269565733477077,"avg_score":0.09269565733477077,"avg_score_percent":9.269565733477076,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.77,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":169121874.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2307577,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":33,"runtime":102235.1495168395,"source_flags":[],"source_harness":"openclaw","source_row_index":570,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":27,"tokens":null,"totalCostUsd":89.77,"totalDurationS":102235.1495168395,"totalInputTokens":169121874.16666666,"totalOutputTokens":2307577,"total_cost_usd":89.77,"total_input_tokens":169121874.16666666,"total_output_tokens":2307577,"total_runtime_s":102235.1495168395,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.269565733477076,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5c66cb74c4c5f3400117","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0c6b2cedb00c508d2b33f504","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07652873704683727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":236,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[117];split=full/full-spectrum;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":236,"metadata":{"avgScore":0.07652873704683727,"avg_score":0.07652873704683727,"avg_score_percent":7.652873704683727,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":39.58,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":61910310.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1024866.5,"passRate":0.03636363636363636,"pass_rate":0.03636363636363636,"pass_rate_percent":3.6363636363636362,"passes":2,"runs":52,"runtime":24283.38227561186,"source_flags":[],"source_harness":"codex","source_row_index":117,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":39.58,"totalDurationS":24283.38227561186,"totalInputTokens":61910310.5,"totalOutputTokens":1024866.5,"total_cost_usd":39.58,"total_input_tokens":61910310.5,"total_output_tokens":1024866.5,"total_runtime_s":24283.38227561186,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.652873704683727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-26c1c98b72e8195dcb38","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0a10d85ce71fd6bc50841ad3","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.06738161075152099,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1072,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[535];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1072,"metadata":{"avgScore":0.06738161075152099,"avg_score":0.06738161075152099,"avg_score_percent":6.7381610751520995,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":24,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":28237254.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":805350.5,"passRate":0.04,"pass_rate":0.04,"pass_rate_percent":4.0,"passes":2,"runs":47,"runtime":16650.077705071075,"source_flags":[],"source_harness":"codex","source_row_index":535,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":45,"tokens":null,"totalCostUsd":24,"totalDurationS":16650.077705071075,"totalInputTokens":28237254.5,"totalOutputTokens":805350.5,"total_cost_usd":24,"total_input_tokens":28237254.5,"total_output_tokens":805350.5,"total_runtime_s":16650.077705071075,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.7381610751520995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b37a258e3c18f9b6a27d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-077dc0f94f71fe086524af8c","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3485832185191737,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1018,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[508];split=unlicensed/full-spectrum;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1018,"metadata":{"avgScore":0.3485832185191737,"avg_score":0.3485832185191737,"avg_score_percent":34.85832185191737,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":92.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":174969360.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":2447842.6666666665,"passRate":0.19333333333333333,"pass_rate":0.19333333333333333,"pass_rate_percent":19.333333333333332,"passes":12,"runs":66,"runtime":145051.03629214008,"source_flags":[],"source_harness":"openclaw","source_row_index":508,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":92.85,"totalDurationS":145051.03629214008,"totalInputTokens":174969360.16666666,"totalOutputTokens":2447842.6666666665,"total_cost_usd":92.85,"total_input_tokens":174969360.16666666,"total_output_tokens":2447842.6666666665,"total_runtime_s":145051.03629214008,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":34.85832185191737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-617a50aa5044f7f7e16a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-04958cf1223e94ce8d86f9c0","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3732607831603129,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":550,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[274];split=full/overall;harness=openclaw;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":550,"metadata":{"avgScore":0.3732607831603129,"avg_score":0.3732607831603129,"avg_score_percent":37.32607831603129,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":274.1,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":488753872.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":7283825.166666667,"passRate":0.2050438596491228,"pass_rate":0.2050438596491228,"pass_rate_percent":20.50438596491228,"passes":40,"runs":179,"runtime":459960.9185098016,"source_flags":[],"source_harness":"openclaw","source_row_index":274,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":274.1,"totalDurationS":459960.9185098016,"totalInputTokens":488753872.1666667,"totalOutputTokens":7283825.166666667,"total_cost_usd":274.1,"total_input_tokens":488753872.1666667,"total_output_tokens":7283825.166666667,"total_runtime_s":459960.9185098016,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.32607831603129,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ff7f2cfd0e50a719a792","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-01aabe3a203ab920ae33eb44","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db8ace0826fa9a860abbc6afebb3cce6c0dc4572","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":354,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[176];split=full/last-exam;harness=codex;model=gpt-5-4;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":354,"metadata":{"avgScore":0,"avg_score":0,"avg_score_percent":0.0,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":45.9,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":66463287.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-4","output_tokens":1142116.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":34,"runtime":43037.737046774244,"source_flags":[],"source_harness":"codex","source_row_index":176,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":45.9,"totalDurationS":43037.737046774244,"totalInputTokens":66463287.5,"totalOutputTokens":1142116.5,"total_cost_usd":45.9,"total_input_tokens":66463287.5,"total_output_tokens":1142116.5,"total_runtime_s":43037.737046774244,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e47e0dc0e8e04751e64b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-fdc3c3c41705ae8235eb50e5","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":115,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[57];split=full;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":115,"metadata":{"avgScore":0.133367903588527,"avg_score":0.133367903588527,"avg_score_percent":13.3367903588527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":134.93,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":206835017.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":3239534.5,"passRate":0.07236842105263158,"pass_rate":0.07236842105263158,"pass_rate_percent":7.236842105263158,"passes":11,"runs":144,"runtime":147936.49156970065,"source_flags":[],"source_harness":"codex","source_row_index":57,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":134.93,"totalDurationS":147936.49156970065,"totalInputTokens":206835017.5,"totalOutputTokens":3239534.5,"total_cost_usd":134.93,"total_input_tokens":206835017.5,"total_output_tokens":3239534.5,"total_runtime_s":147936.49156970065,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.236842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dfb3eb34ec8c139cd645","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fc4820c233a58428d9c3d982","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1141,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[570];split=unlicensed/last-exam;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1141,"metadata":{"avgScore":0.09269565733477077,"avg_score":0.09269565733477077,"avg_score_percent":9.269565733477076,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.77,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":169121874.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2307577,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":33,"runtime":102235.1495168395,"source_flags":[],"source_harness":"openclaw","source_row_index":570,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":27,"tokens":null,"totalCostUsd":89.77,"totalDurationS":102235.1495168395,"totalInputTokens":169121874.16666666,"totalOutputTokens":2307577,"total_cost_usd":89.77,"total_input_tokens":169121874.16666666,"total_output_tokens":2307577,"total_runtime_s":102235.1495168395,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d43ec4a256d882ed953e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fc3c4b2c088ab654a0bae493","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":697,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[348];split=licensed;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":697,"metadata":{"avgScore":0.041695254559251674,"avg_score":0.041695254559251674,"avg_score_percent":4.169525455925167,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":32.38,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":70118711,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":504819,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":82357.47953877086,"source_flags":[],"source_harness":"openclaw","source_row_index":348,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":32.38,"totalDurationS":82357.47953877086,"totalInputTokens":70118711,"totalOutputTokens":504819,"total_cost_usd":32.38,"total_input_tokens":70118711,"total_output_tokens":504819,"total_runtime_s":82357.47953877086,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-762c3e0c0dba422c4155","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f743f455a463aed1a7c7b501","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":353,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[176];split=full/last-exam;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":353,"metadata":{"avgScore":0,"avg_score":0,"avg_score_percent":0.0,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":45.9,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":66463287.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1142116.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":34,"runtime":43037.737046774244,"source_flags":[],"source_harness":"codex","source_row_index":176,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":45.9,"totalDurationS":43037.737046774244,"totalInputTokens":66463287.5,"totalOutputTokens":1142116.5,"total_cost_usd":45.9,"total_input_tokens":66463287.5,"total_output_tokens":1142116.5,"total_runtime_s":43037.737046774244,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0c0724c7c916d344a664","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e8435619ab415c1e5e9fec20","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":319,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[159];split=full/last-exam;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":319,"metadata":{"avgScore":0.07318078210639797,"avg_score":0.07318078210639797,"avg_score_percent":7.318078210639797,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":106.76,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":206932842.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2518416,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":37,"runtime":130535.2900615723,"source_flags":[],"source_harness":"openclaw","source_row_index":159,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":106.76,"totalDurationS":130535.2900615723,"totalInputTokens":206932842.16666666,"totalOutputTokens":2518416,"total_cost_usd":106.76,"total_input_tokens":206932842.16666666,"total_output_tokens":2518416,"total_runtime_s":130535.2900615723,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d2d0f6bc80005be8556d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e1137040f1f6d490868fec8b","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13432835820895522,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1305,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[652];split=unlicensed/near-term;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1305,"metadata":{"avgScore":0.23974389265492613,"avg_score":0.23974389265492613,"avg_score_percent":23.97438926549261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":56.14,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":85510462,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1248047,"passRate":0.13432835820895522,"pass_rate":0.13432835820895522,"pass_rate_percent":13.432835820895523,"passes":9,"runs":68,"runtime":84571.9589642135,"source_flags":[],"source_harness":"codex","source_row_index":652,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":56.14,"totalDurationS":84571.9589642135,"totalInputTokens":85510462,"totalOutputTokens":1248047,"total_cost_usd":56.14,"total_input_tokens":85510462,"total_output_tokens":1248047,"total_runtime_s":84571.9589642135,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.432835820895523,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-377abc22354e4bf4f2be","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c6b7074014a87ce756bcc901","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":673,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[336];split=licensed;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":673,"metadata":{"avgScore":0.07636363636363636,"avg_score":0.07636363636363636,"avg_score_percent":7.636363636363636,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":17.87,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":38055264,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":251023,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":26224.622477491153,"source_flags":[],"source_harness":"codex","source_row_index":336,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":17.87,"totalDurationS":26224.622477491153,"totalInputTokens":38055264,"totalOutputTokens":251023,"total_cost_usd":17.87,"total_input_tokens":38055264,"total_output_tokens":251023,"total_runtime_s":26224.622477491153,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1536451182b32015bfc","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bab5088e748db298a31df020","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1129,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[564];split=unlicensed/last-exam;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1129,"metadata":{"avgScore":0.10303376953774213,"avg_score":0.10303376953774213,"avg_score_percent":10.303376953774212,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":91.02,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":296584939,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":472813,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":28,"runtime":56400.74217087712,"source_flags":[],"source_harness":"ale_claw","source_row_index":564,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":91.02,"totalDurationS":56400.74217087712,"totalInputTokens":296584939,"totalOutputTokens":472813,"total_cost_usd":91.02,"total_input_tokens":296584939,"total_output_tokens":472813,"total_runtime_s":56400.74217087712,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8783747c151a25b6d874","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b33c1658fe6d6a198fcd5a1d","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22104018912529552,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":913,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[456];split=unlicensed;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":913,"metadata":{"avgScore":0.39912759744833903,"avg_score":0.39912759744833903,"avg_score_percent":39.912759744833906,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":241.72,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":418635161.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":6779006.166666667,"passRate":0.22104018912529552,"pass_rate":0.22104018912529552,"pass_rate_percent":22.10401891252955,"passes":40,"runs":172,"runtime":377603.43897103076,"source_flags":[],"source_harness":"openclaw","source_row_index":456,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":241.72,"totalDurationS":377603.43897103076,"totalInputTokens":418635161.1666667,"totalOutputTokens":6779006.166666667,"total_cost_usd":241.72,"total_input_tokens":418635161.1666667,"total_output_tokens":6779006.166666667,"total_runtime_s":377603.43897103076,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.10401891252955,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-67d900081c8c0cb22220","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b20dbb5a0d109cc8b3d24d7a","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19393939393939394,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":181,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[90];split=full/full-spectrum;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":181,"metadata":{"avgScore":0.3434147041110992,"avg_score":0.3434147041110992,"avg_score_percent":34.34147041110992,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":122.6,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":239603010.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2920663.6666666665,"passRate":0.19393939393939394,"pass_rate":0.19393939393939394,"pass_rate_percent":19.393939393939394,"passes":13,"runs":71,"runtime":208103.63729115008,"source_flags":[],"source_harness":"openclaw","source_row_index":90,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":122.6,"totalDurationS":208103.63729115008,"totalInputTokens":239603010.16666666,"totalOutputTokens":2920663.6666666665,"total_cost_usd":122.6,"total_input_tokens":239603010.16666666,"total_output_tokens":2920663.6666666665,"total_runtime_s":208103.63729115008,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.393939393939394,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f4aa7093ee3c8f669303","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-adf338934b7ec7c747d07122","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1276595744680851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1405,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[702];split=unlicensed/overall;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1405,"metadata":{"avgScore":0.30617200210871603,"avg_score":0.30617200210871603,"avg_score_percent":30.617200210871605,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":247.31,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":771282392,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1721086,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":137,"runtime":156584.72501084246,"source_flags":[],"source_harness":"ale_claw","source_row_index":702,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":247.31,"totalDurationS":156584.72501084246,"totalInputTokens":771282392,"totalOutputTokens":1721086,"total_cost_usd":247.31,"total_input_tokens":771282392,"total_output_tokens":1721086,"total_runtime_s":156584.72501084246,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.76595744680851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-894d82c9f85e7cba01b2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ad7dceccfe3e9926e6499c7f","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.04,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1071,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[535];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1071,"metadata":{"avgScore":0.06738161075152099,"avg_score":0.06738161075152099,"avg_score_percent":6.7381610751520995,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":24,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":28237254.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":805350.5,"passRate":0.04,"pass_rate":0.04,"pass_rate_percent":4.0,"passes":2,"runs":47,"runtime":16650.077705071075,"source_flags":[],"source_harness":"codex","source_row_index":535,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":45,"tokens":null,"totalCostUsd":24,"totalDurationS":16650.077705071075,"totalInputTokens":28237254.5,"totalOutputTokens":805350.5,"total_cost_usd":24,"total_input_tokens":28237254.5,"total_output_tokens":805350.5,"total_runtime_s":16650.077705071075,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-78bf2c3cbbbe2b0ca5c3","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a90cc8ebc6165e3f3e2ab674","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":313,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[156];split=full/last-exam;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":313,"metadata":{"avgScore":0.08134244963505957,"avg_score":0.08134244963505957,"avg_score_percent":8.134244963505957,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":100.25,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":321376028,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":540717,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":31,"runtime":63550.53847987812,"source_flags":[],"source_harness":"ale_claw","source_row_index":156,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":31,"tokens":null,"totalCostUsd":100.25,"totalDurationS":63550.53847987812,"totalInputTokens":321376028,"totalOutputTokens":540717,"total_cost_usd":100.25,"total_input_tokens":321376028,"total_output_tokens":540717,"total_runtime_s":63550.53847987812,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e59b407a0ae19004a76","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9c56b2b0ec7480e3afde3485","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1276595744680851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":933,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[466];split=unlicensed;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":933,"metadata":{"avgScore":0.30617200210871603,"avg_score":0.30617200210871603,"avg_score_percent":30.617200210871605,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":247.31,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":771282392,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1721086,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":137,"runtime":156584.72501084246,"source_flags":[],"source_harness":"ale_claw","source_row_index":466,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":247.31,"totalDurationS":156584.72501084246,"totalInputTokens":771282392,"totalOutputTokens":1721086,"total_cost_usd":247.31,"total_input_tokens":771282392,"total_output_tokens":1721086,"total_runtime_s":156584.72501084246,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.76595744680851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e0fa7c50395071df20e0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-9bc2109cb2621fd9692a4325","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.03636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":235,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[117];split=full/full-spectrum;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":235,"metadata":{"avgScore":0.07652873704683727,"avg_score":0.07652873704683727,"avg_score_percent":7.652873704683727,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":39.58,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":61910310.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1024866.5,"passRate":0.03636363636363636,"pass_rate":0.03636363636363636,"pass_rate_percent":3.6363636363636362,"passes":2,"runs":52,"runtime":24283.38227561186,"source_flags":[],"source_harness":"codex","source_row_index":117,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":39.58,"totalDurationS":24283.38227561186,"totalInputTokens":61910310.5,"totalOutputTokens":1024866.5,"total_cost_usd":39.58,"total_input_tokens":61910310.5,"total_output_tokens":1024866.5,"total_runtime_s":24283.38227561186,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.6363636363636362,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bf9b7050c3705d492785","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-957e26d2e38c99765fd55892","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3358208955223881,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":435,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[217];split=full/near-term;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":435,"metadata":{"avgScore":0.578007915138166,"avg_score":0.578007915138166,"avg_score_percent":57.8007915138166,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.88,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86038352.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2379425,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":30,"runs":84,"runtime":147335.50355529715,"source_flags":[],"source_harness":"openclaw","source_row_index":217,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":68.88,"totalDurationS":147335.50355529715,"totalInputTokens":86038352.5,"totalOutputTokens":2379425,"total_cost_usd":68.88,"total_input_tokens":86038352.5,"total_output_tokens":2379425,"total_runtime_s":147335.50355529715,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.582089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8e2e65559f9a31d67006","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-946fd2c714e0d996833cd67c","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2050438596491228,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":549,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[274];split=full/overall;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":549,"metadata":{"avgScore":0.3732607831603129,"avg_score":0.3732607831603129,"avg_score_percent":37.32607831603129,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":274.1,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":488753872.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":7283825.166666667,"passRate":0.2050438596491228,"pass_rate":0.2050438596491228,"pass_rate_percent":20.50438596491228,"passes":40,"runs":179,"runtime":459960.9185098016,"source_flags":[],"source_harness":"openclaw","source_row_index":274,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":274.1,"totalDurationS":459960.9185098016,"totalInputTokens":488753872.1666667,"totalOutputTokens":7283825.166666667,"total_cost_usd":274.1,"total_input_tokens":488753872.1666667,"total_output_tokens":7283825.166666667,"total_runtime_s":459960.9185098016,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.50438596491228,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1c0aa8f30dd4aeb0a7f9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8eba3601636670a9c717ae7d","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11842105263157894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":97,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[48];split=full;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":97,"metadata":{"avgScore":0.2863832387982169,"avg_score":0.2863832387982169,"avg_score_percent":28.63832387982169,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":335.05,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1070761845,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2049232,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":143,"runtime":199675.9292053005,"source_flags":[],"source_harness":"ale_claw","source_row_index":48,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":335.05,"totalDurationS":199675.9292053005,"totalInputTokens":1070761845,"totalOutputTokens":2049232,"total_cost_usd":335.05,"total_input_tokens":1070761845,"total_output_tokens":2049232,"total_runtime_s":199675.9292053005,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.842105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7d4a1dcc7490a7d8944a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-88f2e49f1bbf05bac5ef5962","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.208955223880597,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1289,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[644];split=unlicensed/near-term;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1289,"metadata":{"avgScore":0.45373662346275323,"avg_score":0.45373662346275323,"avg_score_percent":45.373662346275324,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":66.17,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":178595579,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":684821,"passRate":0.208955223880597,"pass_rate":0.208955223880597,"pass_rate_percent":20.8955223880597,"passes":14,"runs":66,"runtime":62517.515991543354,"source_flags":[],"source_harness":"ale_claw","source_row_index":644,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":66.17,"totalDurationS":62517.515991543354,"totalInputTokens":178595579,"totalOutputTokens":684821,"total_cost_usd":66.17,"total_input_tokens":178595579,"total_output_tokens":684821,"total_runtime_s":62517.515991543354,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.8955223880597,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b170bb5025aa312716c2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8889bb686f6e198a8d9e0001","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2253968253968254,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":777,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[388];split=linux_only;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":777,"metadata":{"avgScore":0.4184820514506088,"avg_score":0.4184820514506088,"avg_score_percent":41.84820514506088,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":169.13,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":285789922,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":4866454.666666667,"passRate":0.2253968253968254,"pass_rate":0.2253968253968254,"pass_rate_percent":22.53968253968254,"passes":32,"runs":130,"runtime":180443.07324218247,"source_flags":[],"source_harness":"openclaw","source_row_index":388,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":169.13,"totalDurationS":180443.07324218247,"totalInputTokens":285789922,"totalOutputTokens":4866454.666666667,"total_cost_usd":169.13,"total_input_tokens":285789922,"total_output_tokens":4866454.666666667,"total_runtime_s":180443.07324218247,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.53968253968254,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f11f3c547e8eb94a995e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-81225fa42011bd0b99b9d61f","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3358208955223881,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1271,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[635];split=unlicensed/near-term;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1271,"metadata":{"avgScore":0.578007915138166,"avg_score":0.578007915138166,"avg_score_percent":57.8007915138166,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.88,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86038352.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2379425,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":30,"runs":84,"runtime":147335.50355529715,"source_flags":[],"source_harness":"openclaw","source_row_index":635,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":68.88,"totalDurationS":147335.50355529715,"totalInputTokens":86038352.5,"totalOutputTokens":2379425,"total_cost_usd":68.88,"total_input_tokens":86038352.5,"total_output_tokens":2379425,"total_runtime_s":147335.50355529715,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.582089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-254fdc5ca6b4f22da640","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-72dde664fba631aa02df911c","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":801,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[400];split=linux_only;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":801,"metadata":{"avgScore":0.33810777616126964,"avg_score":0.33810777616126964,"avg_score_percent":33.81077761612696,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":168.38,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":541326506,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1131556,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":14,"runs":103,"runtime":95462.72998667358,"source_flags":[],"source_harness":"ale_claw","source_row_index":400,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":168.38,"totalDurationS":95462.72998667358,"totalInputTokens":541326506,"totalOutputTokens":1131556,"total_cost_usd":168.38,"total_input_tokens":541326506,"total_output_tokens":1131556,"total_runtime_s":95462.72998667358,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.333333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b6ddba66e378fdd8bee4","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-72607593fd6cd9434bd72042","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22104018912529552,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1385,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[692];split=unlicensed/overall;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1385,"metadata":{"avgScore":0.39912759744833903,"avg_score":0.39912759744833903,"avg_score_percent":39.912759744833906,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":241.72,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":418635161.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":6779006.166666667,"passRate":0.22104018912529552,"pass_rate":0.22104018912529552,"pass_rate_percent":22.10401891252955,"passes":40,"runs":172,"runtime":377603.43897103076,"source_flags":[],"source_harness":"openclaw","source_row_index":692,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":241.72,"totalDurationS":377603.43897103076,"totalInputTokens":418635161.1666667,"totalOutputTokens":6779006.166666667,"total_cost_usd":241.72,"total_input_tokens":418635161.1666667,"total_output_tokens":6779006.166666667,"total_runtime_s":377603.43897103076,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.10401891252955,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5106c6aa167ee9d216a2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7145684ff69c5626c18ac174","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07801418439716312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1423,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[711];split=unlicensed/overall;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1423,"metadata":{"avgScore":0.13781504500323477,"avg_score":0.13781504500323477,"avg_score_percent":13.781504500323477,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":117.05,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":168779753.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2988511.5,"passRate":0.07801418439716312,"pass_rate":0.07801418439716312,"pass_rate_percent":7.801418439716312,"passes":11,"runs":137,"runtime":121711.8690922095,"source_flags":[],"source_harness":"codex","source_row_index":711,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":117.05,"totalDurationS":121711.8690922095,"totalInputTokens":168779753.5,"totalOutputTokens":2988511.5,"total_cost_usd":117.05,"total_input_tokens":168779753.5,"total_output_tokens":2988511.5,"total_runtime_s":121711.8690922095,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.801418439716312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9608c9306994d4fd76ff","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5cae43f7817a652400fbd72c","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.208955223880597,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":453,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[226];split=full/near-term;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":453,"metadata":{"avgScore":0.45373662346275323,"avg_score":0.45373662346275323,"avg_score_percent":45.373662346275324,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":66.17,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":178595579,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":684821,"passRate":0.208955223880597,"pass_rate":0.208955223880597,"pass_rate_percent":20.8955223880597,"passes":14,"runs":66,"runtime":62517.515991543354,"source_flags":[],"source_harness":"ale_claw","source_row_index":226,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":66.17,"totalDurationS":62517.515991543354,"totalInputTokens":178595579,"totalOutputTokens":684821,"total_cost_usd":66.17,"total_input_tokens":178595579,"total_output_tokens":684821,"total_runtime_s":62517.515991543354,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.8955223880597,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-da584aa7670ac72a034a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-522467c21e4ad8b983d6e719","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1055,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[527];split=unlicensed/full-spectrum;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1055,"metadata":{"avgScore":0.22512809050649008,"avg_score":0.22512809050649008,"avg_score_percent":22.512809050649008,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":93.03,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":301408364,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":615102,"passRate":0.08,"pass_rate":0.08,"pass_rate_percent":8.0,"passes":4,"runs":49,"runtime":40180.566417631,"source_flags":[],"source_harness":"ale_claw","source_row_index":527,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":93.03,"totalDurationS":40180.566417631,"totalInputTokens":301408364,"totalOutputTokens":615102,"total_cost_usd":93.03,"total_input_tokens":301408364,"total_output_tokens":615102,"total_runtime_s":40180.566417631,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9b902d6ee031fca92b35","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4d2e9e146c54530b799e5a1e","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1189,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[594];split=unlicensed/last-exam;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1189,"metadata":{"avgScore":0,"avg_score":0,"avg_score_percent":0.0,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":40.25,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":58038584.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1069865.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":23240.311491120607,"source_flags":[],"source_harness":"codex","source_row_index":594,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":40.25,"totalDurationS":23240.311491120607,"totalInputTokens":58038584.5,"totalOutputTokens":1069865.5,"total_cost_usd":40.25,"total_input_tokens":58038584.5,"total_output_tokens":1069865.5,"total_runtime_s":23240.311491120607,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-893c0e51fb028e17741c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-47c83f827dc25ca188d4f69c","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.06666666666666667,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":835,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[417];split=linux_only;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":835,"metadata":{"avgScore":0.13116963841633533,"avg_score":0.13116963841633533,"avg_score_percent":13.116963841633533,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":62.4,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":85353349,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1849844,"passRate":0.06666666666666667,"pass_rate":0.06666666666666667,"pass_rate_percent":6.666666666666667,"passes":7,"runs":102,"runtime":39510.33435794874,"source_flags":[],"source_harness":"codex","source_row_index":417,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":62.4,"totalDurationS":39510.33435794874,"totalInputTokens":85353349,"totalOutputTokens":1849844,"total_cost_usd":62.4,"total_input_tokens":85353349,"total_output_tokens":1849844,"total_runtime_s":39510.33435794874,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":6.666666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8a332b0bd4c8d9142691","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4232e5da5de7ded6d39e7dd2","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2050438596491228,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":77,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[38];split=full;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":77,"metadata":{"avgScore":0.3732607831603129,"avg_score":0.3732607831603129,"avg_score_percent":37.32607831603129,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":274.1,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":488753872.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":7283825.166666667,"passRate":0.2050438596491228,"pass_rate":0.2050438596491228,"pass_rate_percent":20.50438596491228,"passes":40,"runs":179,"runtime":459960.9185098016,"source_flags":[],"source_harness":"openclaw","source_row_index":38,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":274.1,"totalDurationS":459960.9185098016,"totalInputTokens":488753872.1666667,"totalOutputTokens":7283825.166666667,"total_cost_usd":274.1,"total_input_tokens":488753872.1666667,"total_output_tokens":7283825.166666667,"total_runtime_s":459960.9185098016,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.50438596491228,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-caf24554c36ebc68b46d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3d25436e9aa3ea245b9f7944","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":215,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[107];split=full/full-spectrum;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":215,"metadata":{"avgScore":0.22938917318771823,"avg_score":0.22938917318771823,"avg_score_percent":22.93891731877182,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":178.85,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":595700026,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":929990,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":5,"runs":54,"runtime":81361.152385172,"source_flags":[],"source_harness":"ale_claw","source_row_index":107,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":178.85,"totalDurationS":81361.152385172,"totalInputTokens":595700026,"totalOutputTokens":929990,"total_cost_usd":178.85,"total_input_tokens":595700026,"total_output_tokens":929990,"total_runtime_s":81361.152385172,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-93ee99ea6cfc2deae137","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3c171c58a8b5f7a545fea448","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1017,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[508];split=unlicensed/full-spectrum;harness=openclaw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1017,"metadata":{"avgScore":0.3485832185191737,"avg_score":0.3485832185191737,"avg_score_percent":34.85832185191737,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":92.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":174969360.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2447842.6666666665,"passRate":0.19333333333333333,"pass_rate":0.19333333333333333,"pass_rate_percent":19.333333333333332,"passes":12,"runs":66,"runtime":145051.03629214008,"source_flags":[],"source_harness":"openclaw","source_row_index":508,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":92.85,"totalDurationS":145051.03629214008,"totalInputTokens":174969360.16666666,"totalOutputTokens":2447842.6666666665,"total_cost_usd":92.85,"total_input_tokens":174969360.16666666,"total_output_tokens":2447842.6666666665,"total_runtime_s":145051.03629214008,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.333333333333332,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-46edd25a6fd05166cc2b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3bc968352e5710c9512917cb","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11842105263157894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":569,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[284];split=full/overall;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":569,"metadata":{"avgScore":0.2863832387982169,"avg_score":0.2863832387982169,"avg_score_percent":28.63832387982169,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":335.05,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1070761845,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2049232,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":143,"runtime":199675.9292053005,"source_flags":[],"source_harness":"ale_claw","source_row_index":284,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":143,"tokens":null,"totalCostUsd":335.05,"totalDurationS":199675.9292053005,"totalInputTokens":1070761845,"totalOutputTokens":2049232,"total_cost_usd":335.05,"total_input_tokens":1070761845,"total_output_tokens":2049232,"total_runtime_s":199675.9292053005,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.842105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a4782fac42930d4e32e2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2ce4d62093b087c195e224fe","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07801418439716312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":951,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[475];split=unlicensed;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":951,"metadata":{"avgScore":0.13781504500323477,"avg_score":0.13781504500323477,"avg_score_percent":13.781504500323477,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":117.05,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":168779753.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":2988511.5,"passRate":0.07801418439716312,"pass_rate":0.07801418439716312,"pass_rate_percent":7.801418439716312,"passes":11,"runs":137,"runtime":121711.8690922095,"source_flags":[],"source_harness":"codex","source_row_index":475,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":117.05,"totalDurationS":121711.8690922095,"totalInputTokens":168779753.5,"totalOutputTokens":2988511.5,"total_cost_usd":117.05,"total_input_tokens":168779753.5,"total_output_tokens":2988511.5,"total_runtime_s":121711.8690922095,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.801418439716312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0b967be3a0aa93b00acf","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2296de3e6bbc2b1169e4dc04","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13432835820895522,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":469,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[234];split=full/near-term;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":469,"metadata":{"avgScore":0.23974389265492613,"avg_score":0.23974389265492613,"avg_score_percent":23.97438926549261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":56.14,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":85510462,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":1248047,"passRate":0.13432835820895522,"pass_rate":0.13432835820895522,"pass_rate_percent":13.432835820895523,"passes":9,"runs":68,"runtime":84571.9589642135,"source_flags":[],"source_harness":"codex","source_row_index":234,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":56.14,"totalDurationS":84571.9589642135,"totalInputTokens":85510462,"totalOutputTokens":1248047,"total_cost_usd":56.14,"total_input_tokens":85510462,"total_output_tokens":1248047,"total_runtime_s":84571.9589642135,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.432835820895523,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-59c2507d85ee3c46a313","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-12244711ac2239c4f7a6a675","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":699,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[349];split=licensed;harness=ale_claw;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":699,"metadata":{"avgScore":0.03272727272727272,"avg_score":0.03272727272727272,"avg_score_percent":3.2727272727272725,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":87.74,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":299479453,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":328146,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":6,"runtime":43091.204194458005,"source_flags":[],"source_harness":"ale_claw","source_row_index":349,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":87.74,"totalDurationS":43091.204194458005,"totalInputTokens":299479453,"totalOutputTokens":328146,"total_cost_usd":87.74,"total_input_tokens":299479453,"total_output_tokens":328146,"total_runtime_s":43091.204194458005,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":36,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-57781c678713f4d81cb4","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-01ef282874cf49f1e64ed870","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.4","modelRef":"gpt-5-4","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.07236842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8db27ebab584ebfa5e8b2bef8784f93bee0aca25","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":587,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[293];split=full/overall;harness=codex;model=gpt-5-4;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":587,"metadata":{"avgScore":0.133367903588527,"avg_score":0.133367903588527,"avg_score_percent":13.3367903588527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":134.93,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":206835017.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-4","output_tokens":3239534.5,"passRate":0.07236842105263158,"pass_rate":0.07236842105263158,"pass_rate_percent":7.236842105263158,"passes":11,"runs":144,"runtime":147936.49156970065,"source_flags":[],"source_harness":"codex","source_row_index":293,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":134.93,"totalDurationS":147936.49156970065,"totalInputTokens":206835017.5,"totalOutputTokens":3239534.5,"total_cost_usd":134.93,"total_input_tokens":206835017.5,"total_output_tokens":3239534.5,"total_runtime_s":147936.49156970065,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.236842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e95feafcc065c2174716","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fc565824dc206c4c30070ebf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.410514071141378,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":152,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[75];split=full/full-spectrum;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":152,"metadata":{"avgScore":0.410514071141378,"avg_score":0.410514071141378,"avg_score_percent":41.0514071141378,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":69.9,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":54901456,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":734763,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":54,"runtime":30454.794733498995,"source_flags":[],"source_harness":"ale_claw","source_row_index":75,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":69.9,"totalDurationS":30454.794733498995,"totalInputTokens":54901456,"totalOutputTokens":734763,"total_cost_usd":69.9,"total_input_tokens":54901456,"total_output_tokens":734763,"total_runtime_s":30454.794733498995,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":41.0514071141378,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5d259e6beb1654449c07","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f2d8c37031b04e6cbe1b8103","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21759244620262602,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":622,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[310];split=licensed;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":622,"metadata":{"avgScore":0.21759244620262602,"avg_score":0.21759244620262602,"avg_score_percent":21.759244620262603,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":34.5,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":41313069,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":175293,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":7,"runtime":10839.482776957002,"source_flags":[],"source_harness":"ale_claw","source_row_index":310,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":34.5,"totalDurationS":10839.482776957002,"totalInputTokens":41313069,"totalOutputTokens":175293,"total_cost_usd":34.5,"total_input_tokens":41313069,"total_output_tokens":175293,"total_runtime_s":10839.482776957002,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.759244620262603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4fca22918da9d33d4c32","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ef075b67f125b20d83bfca67","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3266237832837682,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1026,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[512];split=unlicensed/full-spectrum;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1026,"metadata":{"avgScore":0.3266237832837682,"avg_score":0.3266237832837682,"avg_score_percent":32.66237832837682,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.18,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":68495850,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":915984,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":47,"runtime":65810.74335963264,"source_flags":[],"source_harness":"openclaw","source_row_index":512,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":47,"tokens":null,"totalCostUsd":79.18,"totalDurationS":65810.74335963264,"totalInputTokens":68495850,"totalOutputTokens":915984,"total_cost_usd":79.18,"total_input_tokens":68495850,"total_output_tokens":915984,"total_runtime_s":65810.74335963264,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.66237832837682,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a73d163acf2ae3e0a349","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-eea3bdaabea73b15905440c7","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.33988274762503207,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":186,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[92];split=full/full-spectrum;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":186,"metadata":{"avgScore":0.33988274762503207,"avg_score":0.33988274762503207,"avg_score_percent":33.98827476250321,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.67,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":36415887,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":552100,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":52,"runtime":62650.35733736605,"source_flags":[],"source_harness":"cursor_cli","source_row_index":92,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":52,"tokens":null,"totalCostUsd":49.67,"totalDurationS":62650.35733736605,"totalInputTokens":36415887,"totalOutputTokens":552100,"total_cost_usd":49.67,"total_input_tokens":36415887,"total_output_tokens":552100,"total_runtime_s":62650.35733736605,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.98827476250321,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cef8dd5ace7f96a6acc1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-eb82ed0df83c5105d0318805","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.42851304932922624,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":896,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[447];split=unlicensed;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":896,"metadata":{"avgScore":0.42851304932922624,"avg_score":0.42851304932922624,"avg_score_percent":42.85130493292262,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":157.35,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":128616018,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1601327,"passRate":0.23049645390070922,"pass_rate":0.23049645390070922,"pass_rate_percent":23.04964539007092,"passes":33,"runs":136,"runtime":169469.4067363392,"source_flags":[],"source_harness":"cursor_cli","source_row_index":447,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":133,"tokens":null,"totalCostUsd":157.35,"totalDurationS":169469.4067363392,"totalInputTokens":128616018,"totalOutputTokens":1601327,"total_cost_usd":157.35,"total_input_tokens":128616018,"total_output_tokens":1601327,"total_runtime_s":169469.4067363392,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.85130493292262,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-12768afed5ba9c609938","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-eb2a916e0f8f45a619abcf79","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.13388631862928818,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":270,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[134];split=full/last-exam;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":270,"metadata":{"avgScore":0.13388631862928818,"avg_score":0.13388631862928818,"avg_score_percent":13.388631862928818,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":187.64,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":216406972,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1211914,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":103,"runtime":119659.35271642619,"source_flags":[],"source_harness":"codex","source_row_index":134,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":187.64,"totalDurationS":119659.35271642619,"totalInputTokens":216406972,"totalOutputTokens":1211914,"total_cost_usd":187.64,"total_input_tokens":216406972,"total_output_tokens":1211914,"total_runtime_s":119659.35271642619,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.388631862928818,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-72761b1cf61fd1159008","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e99c8276e03ac13b2cf0111e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.22245515154818696,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":616,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[307];split=licensed;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":616,"metadata":{"avgScore":0.22245515154818696,"avg_score":0.22245515154818696,"avg_score_percent":22.245515154818698,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":65.73,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":59208218,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":212510,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":91032,"source_flags":[],"source_harness":"codex","source_row_index":307,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":65.73,"totalDurationS":91032,"totalInputTokens":59208218,"totalOutputTokens":212510,"total_cost_usd":65.73,"total_input_tokens":59208218,"total_output_tokens":212510,"total_runtime_s":91032,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.245515154818698,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1b4aac8fd82d44e0fece","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e7271d47f8c7262cd2ebddb1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6228643923694662,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":410,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[204];split=full/near-term;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":410,"metadata":{"avgScore":0.6228643923694662,"avg_score":0.6228643923694662,"avg_score_percent":62.28643923694662,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.46,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":51827010,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":741097,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":23,"runs":69,"runtime":108587.11174137115,"source_flags":[],"source_harness":"cursor_cli","source_row_index":204,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":68.46,"totalDurationS":108587.11174137115,"totalInputTokens":51827010,"totalOutputTokens":741097,"total_cost_usd":68.46,"total_input_tokens":51827010,"total_output_tokens":741097,"total_runtime_s":108587.11174137115,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":62.28643923694662,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e0022501d742b731e048","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e5e00d7ca31e4bf321855dc6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6736635047623744,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":390,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[194];split=full/near-term;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":390,"metadata":{"avgScore":0.6736635047623744,"avg_score":0.6736635047623744,"avg_score_percent":67.36635047623743,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":147.52,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":167219189,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1025815,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":67,"runtime":53801.491831077,"source_flags":[],"source_harness":"ale_claw","source_row_index":194,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":147.52,"totalDurationS":53801.491831077,"totalInputTokens":167219189,"totalOutputTokens":1025815,"total_cost_usd":147.52,"total_input_tokens":167219189,"total_output_tokens":1025815,"total_runtime_s":53801.491831077,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":67.36635047623743,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e6d70a66c2cd8d4d859a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e5b420cb37337c2295d94c11","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.42851304932922624,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1368,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[683];split=unlicensed/overall;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1368,"metadata":{"avgScore":0.42851304932922624,"avg_score":0.42851304932922624,"avg_score_percent":42.85130493292262,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":157.35,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":128616018,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1601327,"passRate":0.23049645390070922,"pass_rate":0.23049645390070922,"pass_rate_percent":23.04964539007092,"passes":33,"runs":136,"runtime":169469.4067363392,"source_flags":[],"source_harness":"cursor_cli","source_row_index":683,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":133,"tokens":null,"totalCostUsd":157.35,"totalDurationS":169469.4067363392,"totalInputTokens":128616018,"totalOutputTokens":1601327,"total_cost_usd":157.35,"total_input_tokens":128616018,"total_output_tokens":1601327,"total_runtime_s":169469.4067363392,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.85130493292262,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5a8b8bdf259ee5db5b16","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e1e1474536100d13e298f942","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.47931391024591824,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":502,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[250];split=full/overall;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":502,"metadata":{"avgScore":0.47931391024591824,"avg_score":0.47931391024591824,"avg_score_percent":47.931391024591825,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":601.69,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":560480756,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":6366308.5,"passRate":0.26644736842105265,"pass_rate":0.26644736842105265,"pass_rate_percent":26.644736842105267,"passes":80,"runs":299,"runtime":349652.5,"source_flags":[],"source_harness":"codex","source_row_index":250,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":601.69,"totalDurationS":349652.5,"totalInputTokens":560480756,"totalOutputTokens":6366308.5,"total_cost_usd":601.69,"total_input_tokens":560480756,"total_output_tokens":6366308.5,"total_runtime_s":349652.5,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.931391024591825,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-445f75062928f708e68e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e1824372df728e8b48e057a1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5353736949071799,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":442,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[220];split=full/near-term;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":442,"metadata":{"avgScore":0.5353736949071799,"avg_score":0.5353736949071799,"avg_score_percent":53.53736949071799,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":48.16,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":28164467,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":528351,"passRate":0.2537313432835821,"pass_rate":0.2537313432835821,"pass_rate_percent":25.37313432835821,"passes":17,"runs":67,"runtime":36974,"source_flags":[],"source_harness":"codex","source_row_index":220,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":48.16,"totalDurationS":36974,"totalInputTokens":28164467,"totalOutputTokens":528351,"total_cost_usd":48.16,"total_input_tokens":28164467,"total_output_tokens":528351,"total_runtime_s":36974,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.53736949071799,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-933526c5fc3f10e085c0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-e010e3e6467895c173c4f88b","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.34037980623548364,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1020,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[509];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1020,"metadata":{"avgScore":0.34037980623548364,"avg_score":0.34037980623548364,"avg_score_percent":34.037980623548364,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":35.01,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":21829853,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":377771,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":50,"runtime":28603,"source_flags":[],"source_harness":"codex","source_row_index":509,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":35.01,"totalDurationS":28603,"totalInputTokens":21829853,"totalOutputTokens":377771,"total_cost_usd":35.01,"total_input_tokens":21829853,"total_output_tokens":377771,"total_runtime_s":28603,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":34.037980623548364,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d04bcc601378f2b36d56","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-da1c358e07fdee0f65911e42","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.670526220421849,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1230,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[614];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1230,"metadata":{"avgScore":0.670526220421849,"avg_score":0.670526220421849,"avg_score_percent":67.0526220421849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":120.93,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":162729818,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1307423.8333333333,"passRate":0.3905472636815919,"pass_rate":0.3905472636815919,"pass_rate_percent":39.054726368159194,"passes":77,"runs":198,"runtime":97188.19844049853,"source_flags":[],"source_harness":"codex","source_row_index":614,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":120.93,"totalDurationS":97188.19844049853,"totalInputTokens":162729818,"totalOutputTokens":1307423.8333333333,"total_cost_usd":120.93,"total_input_tokens":162729818,"total_output_tokens":1307423.8333333333,"total_runtime_s":97188.19844049853,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.0526220421849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a3403fb5c95a42333b6c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d939cdbb4efdf80f51448a54","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.22151974870087496,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":620,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[309];split=licensed;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":620,"metadata":{"avgScore":0.22151974870087496,"avg_score":0.22151974870087496,"avg_score_percent":22.151974870087496,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":76.54,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":71991083.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":346287,"passRate":0.0303030303030303,"pass_rate":0.0303030303030303,"pass_rate_percent":3.03030303030303,"passes":1,"runs":26,"runtime":44609.951294866376,"source_flags":[],"source_harness":"codex","source_row_index":309,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":76.54,"totalDurationS":44609.951294866376,"totalInputTokens":71991083.5,"totalOutputTokens":346287,"total_cost_usd":76.54,"total_input_tokens":71991083.5,"total_output_tokens":346287,"total_runtime_s":44609.951294866376,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.151974870087496,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5a0ebb5540757d504872","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d5332267a38fc59c01057542","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.49867589404218615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1338,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[668];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1338,"metadata":{"avgScore":0.49867589404218615,"avg_score":0.49867589404218615,"avg_score_percent":49.86758940421861,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":522.76,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":481345798.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":5745106.5,"passRate":0.2872340425531915,"pass_rate":0.2872340425531915,"pass_rate_percent":28.723404255319153,"passes":80,"runs":278,"runtime":287393.5,"source_flags":[],"source_harness":"codex","source_row_index":668,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":522.76,"totalDurationS":287393.5,"totalInputTokens":481345798.5,"totalOutputTokens":5745106.5,"total_cost_usd":522.76,"total_input_tokens":481345798.5,"total_output_tokens":5745106.5,"total_runtime_s":287393.5,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.86758940421861,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6b66f12232adf74fc9c9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d373ed3ffa206ca708fbd1c5","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.44245775446794344,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1358,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[678];split=unlicensed/overall;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1358,"metadata":{"avgScore":0.44245775446794344,"avg_score":0.44245775446794344,"avg_score_percent":44.24577544679435,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":394.14,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":403857647,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3083904,"passRate":0.23404255319148937,"pass_rate":0.23404255319148937,"pass_rate_percent":23.404255319148938,"passes":33,"runs":131,"runtime":273966.4079745095,"source_flags":[],"source_harness":"openclaw","source_row_index":678,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":131,"tokens":null,"totalCostUsd":394.14,"totalDurationS":273966.4079745095,"totalInputTokens":403857647,"totalOutputTokens":3083904,"total_cost_usd":394.14,"total_input_tokens":403857647,"total_output_tokens":3083904,"total_runtime_s":273966.4079745095,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":44.24577544679435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f3b8084bdd0a19bc2291","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d3334f1c81b648c6e8e0779d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4605257773323447,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":748,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[373];split=linux_only;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":748,"metadata":{"avgScore":0.4605257773323447,"avg_score":0.4605257773323447,"avg_score_percent":46.052577733234465,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":218.69,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":211187431,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2152504,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":101,"runtime":113408.57987576415,"source_flags":[],"source_harness":"openclaw","source_row_index":373,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":218.69,"totalDurationS":113408.57987576415,"totalInputTokens":211187431,"totalOutputTokens":2152504,"total_cost_usd":218.69,"total_input_tokens":211187431,"total_output_tokens":2152504,"total_runtime_s":113408.57987576415,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":46.052577733234465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-78646436fa62f140c2ce","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d2f7dac353aef25697087f12","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3636965114654929,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":80,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[39];split=full;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":80,"metadata":{"avgScore":0.3636965114654929,"avg_score":0.3636965114654929,"avg_score_percent":36.369651146549295,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":146.61,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":109294354,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1270291,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":139211,"source_flags":[],"source_harness":"codex","source_row_index":39,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":146.61,"totalDurationS":139211,"totalInputTokens":109294354,"totalOutputTokens":1270291,"total_cost_usd":146.61,"total_input_tokens":109294354,"total_output_tokens":1270291,"total_runtime_s":139211,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.369651146549295,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0ab993e5e94903c4857f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d0f2695fd76ae814674e3bea","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.42072396960513175,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":52,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[25];split=full;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":52,"metadata":{"avgScore":0.42072396960513175,"avg_score":0.42072396960513175,"avg_score_percent":42.072396960513174,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446.52,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":469259907,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3279175,"passRate":0.21710526315789475,"pass_rate":0.21710526315789475,"pass_rate_percent":21.710526315789476,"passes":33,"runs":136,"runtime":297468.11566528457,"source_flags":[],"source_harness":"openclaw","source_row_index":25,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":136,"tokens":null,"totalCostUsd":446.52,"totalDurationS":297468.11566528457,"totalInputTokens":469259907,"totalOutputTokens":3279175,"total_cost_usd":446.52,"total_input_tokens":469259907,"total_output_tokens":3279175,"total_runtime_s":297468.11566528457,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.072396960513174,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-042de638982f9bc55fd6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ce19c29f6eb70093573e9a7c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3361494333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1024,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[511];split=unlicensed/full-spectrum;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1024,"metadata":{"avgScore":0.3361494333333333,"avg_score":0.3361494333333333,"avg_score_percent":33.61494333333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.52,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":31812048,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":496590,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":10,"runs":48,"runtime":42562.361035950235,"source_flags":[],"source_harness":"cursor_cli","source_row_index":511,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":44.52,"totalDurationS":42562.361035950235,"totalInputTokens":31812048,"totalOutputTokens":496590,"total_cost_usd":44.52,"total_input_tokens":31812048,"total_output_tokens":496590,"total_runtime_s":42562.361035950235,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.61494333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2961fa09f8c95b127a39","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cc49b0e5c6460c7b9c6812ec","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.44245775446794344,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":886,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[442];split=unlicensed;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":886,"metadata":{"avgScore":0.44245775446794344,"avg_score":0.44245775446794344,"avg_score_percent":44.24577544679435,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":394.14,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":403857647,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3083904,"passRate":0.23404255319148937,"pass_rate":0.23404255319148937,"pass_rate_percent":23.404255319148938,"passes":33,"runs":131,"runtime":273966.4079745095,"source_flags":[],"source_harness":"openclaw","source_row_index":442,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":131,"tokens":null,"totalCostUsd":394.14,"totalDurationS":273966.4079745095,"totalInputTokens":403857647,"totalOutputTokens":3083904,"total_cost_usd":394.14,"total_input_tokens":403857647,"total_output_tokens":3083904,"total_runtime_s":273966.4079745095,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":44.24577544679435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1ec2dc88175fdaf07bd9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cbdbd2cc0626cbeb0d68ce30","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12815059202118556,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":274,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[136];split=full/last-exam;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":274,"metadata":{"avgScore":0.12815059202118556,"avg_score":0.12815059202118556,"avg_score_percent":12.815059202118556,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.32,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":126287850,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":704360,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":32,"runtime":49458.08196799901,"source_flags":[],"source_harness":"ale_claw","source_row_index":136,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":107.32,"totalDurationS":49458.08196799901,"totalInputTokens":126287850,"totalOutputTokens":704360,"total_cost_usd":107.32,"total_input_tokens":126287850,"total_output_tokens":704360,"total_runtime_s":49458.08196799901,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.815059202118556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fba9cf2ddfb1642f67ec","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cb347015ced34a9b3308dfc1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3828945809408762,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":916,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[457];split=unlicensed;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":916,"metadata":{"avgScore":0.3828945809408762,"avg_score":0.3828945809408762,"avg_score_percent":38.28945809408762,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.9,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":96188305,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1176916,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":108014,"source_flags":[],"source_harness":"codex","source_row_index":457,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":127.9,"totalDurationS":108014,"totalInputTokens":96188305,"totalOutputTokens":1176916,"total_cost_usd":127.9,"total_input_tokens":96188305,"total_output_tokens":1176916,"total_runtime_s":108014,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.28945809408762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1032e5f3d7f72773900","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ca2eccd36602eec862317ef1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6719239733700241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1228,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[613];split=unlicensed/near-term;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1228,"metadata":{"avgScore":0.6719239733700241,"avg_score":0.6719239733700241,"avg_score_percent":67.1923973370024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":218.12,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":223015549,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1462808,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":148769.07770477302,"source_flags":[],"source_harness":"openclaw","source_row_index":613,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":218.12,"totalDurationS":148769.07770477302,"totalInputTokens":223015549,"totalOutputTokens":1462808,"total_cost_usd":218.12,"total_input_tokens":223015549,"total_output_tokens":1462808,"total_runtime_s":148769.07770477302,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":67.1923973370024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-df72e2767896f2c35f16","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c8a0974f156b98b8d8cf406e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.44715784362305094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":514,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[256];split=full/overall;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":514,"metadata":{"avgScore":0.44715784362305094,"avg_score":0.44715784362305094,"avg_score_percent":44.715784362305094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.59,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":489884949,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3517360.8333333335,"passRate":0.2445175438596491,"pass_rate":0.2445175438596491,"pass_rate_percent":24.45175438596491,"passes":110,"runs":438,"runtime":270900.40022482676,"source_flags":[],"source_harness":"codex","source_row_index":256,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":415.59,"totalDurationS":270900.40022482676,"totalInputTokens":489884949,"totalOutputTokens":3517360.8333333335,"total_cost_usd":415.59,"total_input_tokens":489884949,"total_output_tokens":3517360.8333333335,"total_runtime_s":270900.40022482676,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.715784362305094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fb87a2f007201448a186","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c8998e48c3365a4dbaca0e91","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.581507678464108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1268,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[633];split=unlicensed/near-term;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1268,"metadata":{"avgScore":0.581507678464108,"avg_score":0.581507678464108,"avg_score_percent":58.1507678464108,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":105.98,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98383491,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":962554,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":65,"runtime":84340.52837546298,"source_flags":[],"source_harness":"droid","source_row_index":633,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":105.98,"totalDurationS":84340.52837546298,"totalInputTokens":98383491,"totalOutputTokens":962554,"total_cost_usd":105.98,"total_input_tokens":98383491,"total_output_tokens":962554,"total_runtime_s":84340.52837546298,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":58.1507678464108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ba0047b3aacdc320f74e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-be908b3641a52f6ee26e4a71","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.09880317382248731,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":306,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[152];split=full/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":306,"metadata":{"avgScore":0.09880317382248731,"avg_score":0.09880317382248731,"avg_score_percent":9.88031738224873,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":64.34,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":59361487,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":378037,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":57793,"source_flags":[],"source_harness":"codex","source_row_index":152,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":64.34,"totalDurationS":57793,"totalInputTokens":59361487,"totalOutputTokens":378037,"total_cost_usd":64.34,"total_input_tokens":59361487,"total_output_tokens":378037,"total_runtime_s":57793,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.88031738224873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7e40a24d44b1b3e9ef5a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bb5f3d1d2d93c33762880940","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4114434329247684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1380,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[689];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1380,"metadata":{"avgScore":0.4114434329247684,"avg_score":0.4114434329247684,"avg_score_percent":41.14434329247684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":291.95,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":235196044,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2640977,"passRate":0.19858156028368795,"pass_rate":0.19858156028368795,"pass_rate_percent":19.858156028368796,"passes":28,"runs":138,"runtime":179609,"source_flags":[],"source_harness":"codex","source_row_index":689,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":291.95,"totalDurationS":179609,"totalInputTokens":235196044,"totalOutputTokens":2640977,"total_cost_usd":291.95,"total_input_tokens":235196044,"total_output_tokens":2640977,"total_runtime_s":179609,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.14434329247684,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f82f5d269839ad619d28","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b9f4928ba8c4728e50e4459a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.39776664940409484,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":68,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[33];split=full;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":68,"metadata":{"avgScore":0.39776664940409484,"avg_score":0.39776664940409484,"avg_score_percent":39.77666494040948,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":357.68,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":294404262,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2853487,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":149,"runtime":270641,"source_flags":[],"source_harness":"codex","source_row_index":33,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":357.68,"totalDurationS":270641,"totalInputTokens":294404262,"totalOutputTokens":2853487,"total_cost_usd":357.68,"total_input_tokens":294404262,"total_output_tokens":2853487,"total_runtime_s":270641,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.77666494040948,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-37fceb8f3550a3b63efc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b3bcee673822acca459254ef","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.14213636363636362,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":650,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[324];split=licensed;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":650,"metadata":{"avgScore":0.14213636363636362,"avg_score":0.14213636363636362,"avg_score_percent":14.213636363636361,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":52.38,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":65402260,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":195271,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":23501.70769077503,"source_flags":[],"source_harness":"openclaw","source_row_index":324,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":52.38,"totalDurationS":23501.70769077503,"totalInputTokens":65402260,"totalOutputTokens":195271,"total_cost_usd":52.38,"total_input_tokens":65402260,"total_output_tokens":195271,"total_runtime_s":23501.70769077503,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.213636363636361,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e8b8264646e6e1dad40","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b1f0033dd8f63ed0d789b611","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4578443243363864,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":508,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[253];split=full/overall;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":508,"metadata":{"avgScore":0.4578443243363864,"avg_score":0.4578443243363864,"avg_score_percent":45.78443243363864,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":309.77,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":334305043,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2340693,"passRate":0.23026315789473684,"pass_rate":0.23026315789473684,"pass_rate_percent":23.026315789473685,"passes":35,"runs":145,"runtime":128044.22844149194,"source_flags":[],"source_harness":"ale_claw","source_row_index":253,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":145,"tokens":null,"totalCostUsd":309.77,"totalDurationS":128044.22844149194,"totalInputTokens":334305043,"totalOutputTokens":2340693,"total_cost_usd":309.77,"total_input_tokens":334305043,"total_output_tokens":2340693,"total_runtime_s":128044.22844149194,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.78443243363864,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a969261b52bd62bdbeea","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b07dca9cf28b15acb0ebb134","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.1336927248629258,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1106,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[552];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1106,"metadata":{"avgScore":0.1336927248629258,"avg_score":0.1336927248629258,"avg_score_percent":13.36927248629258,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":135.97,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":169216522.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":987069.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":85,"runtime":87267.92330068085,"source_flags":[],"source_harness":"codex","source_row_index":552,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":135.97,"totalDurationS":87267.92330068085,"totalInputTokens":169216522.5,"totalOutputTokens":987069.5,"total_cost_usd":135.97,"total_input_tokens":169216522.5,"total_output_tokens":987069.5,"total_runtime_s":87267.92330068085,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.36927248629258,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-43d71bb99fa0697550af","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b025afdec8c791b96a2e38f7","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.160932671435888,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":258,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[128];split=full/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":258,"metadata":{"avgScore":0.160932671435888,"avg_score":0.160932671435888,"avg_score_percent":16.0932671435888,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":283.24,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":316923871.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2234417.5,"passRate":0.013157894736842105,"pass_rate":0.013157894736842105,"pass_rate_percent":1.3157894736842104,"passes":1,"runs":74,"runtime":153534.5,"source_flags":[],"source_harness":"codex","source_row_index":128,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":283.24,"totalDurationS":153534.5,"totalInputTokens":316923871.5,"totalOutputTokens":2234417.5,"total_cost_usd":283.24,"total_input_tokens":316923871.5,"total_output_tokens":2234417.5,"total_runtime_s":153534.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.0932671435888,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-245bdee9fd23d1fda548","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ae811d30a9ffa294c9d99893","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.581507678464108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":432,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[215];split=full/near-term;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":432,"metadata":{"avgScore":0.581507678464108,"avg_score":0.581507678464108,"avg_score_percent":58.1507678464108,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":105.98,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98383491,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":962554,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":65,"runtime":84340.52837546298,"source_flags":[],"source_harness":"droid","source_row_index":215,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":105.98,"totalDurationS":84340.52837546298,"totalInputTokens":98383491,"totalOutputTokens":962554,"total_cost_usd":105.98,"total_input_tokens":98383491,"total_output_tokens":962554,"total_runtime_s":84340.52837546298,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":58.1507678464108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5d428eaddeec2de6b1c4","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ad4476988e7382c15e1adf5d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.43427017991504974,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":766,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[382];split=linux_only;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":766,"metadata":{"avgScore":0.43427017991504974,"avg_score":0.43427017991504974,"avg_score_percent":43.427017991504975,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":97.48,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":68450446,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1079925,"passRate":0.22857142857142856,"pass_rate":0.22857142857142856,"pass_rate_percent":22.857142857142858,"passes":24,"runs":100,"runtime":52191.603746200475,"source_flags":[],"source_harness":"cursor_cli","source_row_index":382,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":100,"tokens":null,"totalCostUsd":97.48,"totalDurationS":52191.603746200475,"totalInputTokens":68450446,"totalOutputTokens":1079925,"total_cost_usd":97.48,"total_input_tokens":68450446,"total_output_tokens":1079925,"total_runtime_s":52191.603746200475,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":43.427017991504975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6ac53c4d3c3c95eef7bf","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aa740a90b817f0e6a1b98274","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6736635047623744,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1226,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[612];split=unlicensed/near-term;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1226,"metadata":{"avgScore":0.6736635047623744,"avg_score":0.6736635047623744,"avg_score_percent":67.36635047623743,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":147.52,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":167219189,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1025815,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":67,"runtime":53801.491831077,"source_flags":[],"source_harness":"ale_claw","source_row_index":612,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":147.52,"totalDurationS":53801.491831077,"totalInputTokens":167219189,"totalOutputTokens":1025815,"total_cost_usd":147.52,"total_input_tokens":167219189,"total_output_tokens":1025815,"total_runtime_s":53801.491831077,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":67.36635047623743,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cac7d520f5136f16d021","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aa0c92796633712391d3331c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3924651258147475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":72,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[35];split=full;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":72,"metadata":{"avgScore":0.3924651258147475,"avg_score":0.3924651258147475,"avg_score_percent":39.246512581474754,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":242.24,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":234782616,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2196372,"passRate":0.19736842105263158,"pass_rate":0.19736842105263158,"pass_rate_percent":19.736842105263158,"passes":30,"runs":139,"runtime":229751.49587275658,"source_flags":[],"source_harness":"droid","source_row_index":35,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":242.24,"totalDurationS":229751.49587275658,"totalInputTokens":234782616,"totalOutputTokens":2196372,"total_cost_usd":242.24,"total_input_tokens":234782616,"total_output_tokens":2196372,"total_runtime_s":229751.49587275658,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":39.246512581474754,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9e1b2ecb083c3c80524f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a9ca4179d5fd4e8205132634","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3519490303044795,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":178,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[88];split=full/full-spectrum;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":178,"metadata":{"avgScore":0.3519490303044795,"avg_score":0.3519490303044795,"avg_score_percent":35.194903030447946,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.85,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":58598314,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":792157,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":50,"runtime":56202.44506297966,"source_flags":[],"source_harness":"droid","source_row_index":88,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":68.85,"totalDurationS":56202.44506297966,"totalInputTokens":58598314,"totalOutputTokens":792157,"total_cost_usd":68.85,"total_input_tokens":58598314,"total_output_tokens":792157,"total_runtime_s":56202.44506297966,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.194903030447946,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-68230d80da3cd1297d9e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a832ec95c99b22ea0feb6462","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5926809886791707,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1266,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[632];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1266,"metadata":{"avgScore":0.5926809886791707,"avg_score":0.5926809886791707,"avg_score_percent":59.268098867917075,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.43,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":81684093,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1064184,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":66,"runtime":48118,"source_flags":[],"source_harness":"codex","source_row_index":632,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":110.43,"totalDurationS":48118,"totalInputTokens":81684093,"totalOutputTokens":1064184,"total_cost_usd":110.43,"total_input_tokens":81684093,"total_output_tokens":1064184,"total_runtime_s":48118,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":59.268098867917075,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-af99349ddc664371e57a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a7ad23ff8cbe268e70f7a290","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.47658737865887824,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1344,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[671];split=unlicensed/overall;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1344,"metadata":{"avgScore":0.47658737865887824,"avg_score":0.47658737865887824,"avg_score_percent":47.65873786588782,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":275.26,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":292991974,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2165400,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":138,"runtime":117204.74566453496,"source_flags":[],"source_harness":"ale_claw","source_row_index":671,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":275.26,"totalDurationS":117204.74566453496,"totalInputTokens":292991974,"totalOutputTokens":2165400,"total_cost_usd":275.26,"total_input_tokens":292991974,"total_output_tokens":2165400,"total_runtime_s":117204.74566453496,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":47.65873786588782,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1bd971dc104601d867d8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a64b9a4613387d74f42bb39e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.333952530257971,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":188,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[93];split=full/full-spectrum;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":188,"metadata":{"avgScore":0.333952530257971,"avg_score":0.333952530257971,"avg_score_percent":33.3952530257971,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":101.27,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98557834,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1053767,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":50,"runtime":82657.90435112063,"source_flags":[],"source_harness":"openclaw","source_row_index":93,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":101.27,"totalDurationS":82657.90435112063,"totalInputTokens":98557834,"totalOutputTokens":1053767,"total_cost_usd":101.27,"total_input_tokens":98557834,"total_output_tokens":1053767,"total_runtime_s":82657.90435112063,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.3952530257971,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-839b2ed21d783993ca60","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a3ecce0130acb4e60461c004","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3503439333349274,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1016,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[507];split=unlicensed/full-spectrum;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1016,"metadata":{"avgScore":0.3503439333349274,"avg_score":0.3503439333349274,"avg_score_percent":35.034393333492744,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.39,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":49922490,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":723508,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":47,"runtime":53747.939585612694,"source_flags":[],"source_harness":"droid","source_row_index":507,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":47,"tokens":null,"totalCostUsd":61.39,"totalDurationS":53747.939585612694,"totalInputTokens":49922490,"totalOutputTokens":723508,"total_cost_usd":61.39,"total_input_tokens":49922490,"total_output_tokens":723508,"total_runtime_s":53747.939585612694,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.034393333492744,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c0d4db7fb86e501e1de6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a3c47bea75d7dfd6766d18bd","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.40680078557980925,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":60,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[29];split=full;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":60,"metadata":{"avgScore":0.40680078557980925,"avg_score":0.40680078557980925,"avg_score_percent":40.68007855798093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":173.52,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":148061075,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1687764,"passRate":0.2138157894736842,"pass_rate":0.2138157894736842,"pass_rate_percent":21.38157894736842,"passes":33,"runs":142,"runtime":211627.30808669614,"source_flags":[],"source_harness":"cursor_cli","source_row_index":29,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":139,"tokens":null,"totalCostUsd":173.52,"totalDurationS":211627.30808669614,"totalInputTokens":148061075,"totalOutputTokens":1687764,"total_cost_usd":173.52,"total_input_tokens":148061075,"total_output_tokens":1687764,"total_runtime_s":211627.30808669614,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":40.68007855798093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8e878e456c882eff8416","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a20892d3c24f2cc04be2ab4f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10732247368421054,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":296,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[147];split=full/last-exam;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":296,"metadata":{"avgScore":0.10732247368421054,"avg_score":0.10732247368421054,"avg_score_percent":10.732247368421053,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.49,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":63892750,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":456416,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":29,"runtime":61917.92565060669,"source_flags":[],"source_harness":"cursor_cli","source_row_index":147,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":29,"tokens":null,"totalCostUsd":61.49,"totalDurationS":61917.92565060669,"totalInputTokens":63892750,"totalOutputTokens":456416,"total_cost_usd":61.49,"total_input_tokens":63892750,"total_output_tokens":456416,"total_runtime_s":61917.92565060669,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.732247368421053,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-72003549860fed79aec0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a1cfe013d8d12c66703d4375","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.46476081556733406,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":878,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[438];split=unlicensed;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":878,"metadata":{"avgScore":0.46476081556733406,"avg_score":0.46476081556733406,"avg_score_percent":46.476081556733405,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":339.05,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":417893865.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3171073.8333333335,"passRate":0.26122931442080377,"pass_rate":0.26122931442080377,"pass_rate_percent":26.12293144208038,"passes":109,"runs":412,"runtime":226290.44892996037,"source_flags":[],"source_harness":"codex","source_row_index":438,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":339.05,"totalDurationS":226290.44892996037,"totalInputTokens":417893865.5,"totalOutputTokens":3171073.8333333335,"total_cost_usd":339.05,"total_input_tokens":417893865.5,"total_output_tokens":3171073.8333333335,"total_runtime_s":226290.44892996037,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.476081556733405,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-60656b9b6fdecd5eb357","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a07525cfcf82d4be04630f63","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.09288872936875796,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1140,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[569];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1140,"metadata":{"avgScore":0.09288872936875796,"avg_score":0.09288872936875796,"avg_score_percent":9.288872936875796,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":48.7,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":48327687,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":310243,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":46146,"source_flags":[],"source_harness":"codex","source_row_index":569,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":48.7,"totalDurationS":46146,"totalInputTokens":48327687,"totalOutputTokens":310243,"total_cost_usd":48.7,"total_input_tokens":48327687,"total_output_tokens":310243,"total_runtime_s":46146,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.288872936875796,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f392b1889d8beec331aa","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9f7da013f5d7cf4962ed5107","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.40680078557980925,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":532,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[265];split=full/overall;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":532,"metadata":{"avgScore":0.40680078557980925,"avg_score":0.40680078557980925,"avg_score_percent":40.68007855798093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":173.52,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":148061075,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1687764,"passRate":0.2138157894736842,"pass_rate":0.2138157894736842,"pass_rate_percent":21.38157894736842,"passes":33,"runs":142,"runtime":211627.30808669614,"source_flags":[],"source_harness":"cursor_cli","source_row_index":265,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":139,"tokens":null,"totalCostUsd":173.52,"totalDurationS":211627.30808669614,"totalInputTokens":148061075,"totalOutputTokens":1687764,"total_cost_usd":173.52,"total_input_tokens":148061075,"total_output_tokens":1687764,"total_runtime_s":211627.30808669614,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":40.68007855798093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8d918e875ddc7c11ee56","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-993d13d5042229c922b9a26c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10874277625057213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":294,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[146];split=full/last-exam;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":294,"metadata":{"avgScore":0.10874277625057213,"avg_score":0.10874277625057213,"avg_score_percent":10.874277625057212,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":154.51,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":178227464,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":996954,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":29,"runtime":89528.58231328387,"source_flags":[],"source_harness":"openclaw","source_row_index":146,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":29,"tokens":null,"totalCostUsd":154.51,"totalDurationS":89528.58231328387,"totalInputTokens":178227464,"totalOutputTokens":996954,"total_cost_usd":154.51,"total_input_tokens":178227464,"total_output_tokens":996954,"total_runtime_s":89528.58231328387,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.874277625057212,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-39956ea12b7f7d8dc3af","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9786bc00f990308b8969f794","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6228643923694662,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1246,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[622];split=unlicensed/near-term;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1246,"metadata":{"avgScore":0.6228643923694662,"avg_score":0.6228643923694662,"avg_score_percent":62.28643923694662,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.46,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":51827010,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":741097,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":23,"runs":69,"runtime":108587.11174137115,"source_flags":[],"source_harness":"cursor_cli","source_row_index":622,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":68.46,"totalDurationS":108587.11174137115,"totalInputTokens":51827010,"totalOutputTokens":741097,"total_cost_usd":68.46,"total_input_tokens":51827010,"total_output_tokens":741097,"total_runtime_s":108587.11174137115,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":62.28643923694662,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dc9b5b3d05231c0dc4fc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9658eacaaaf67bfe2f9d93de","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4136390008783094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1378,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[688];split=unlicensed/overall;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1378,"metadata":{"avgScore":0.4136390008783094,"avg_score":0.4136390008783094,"avg_score_percent":41.36390008783094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":224.97,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":214695571,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2091602,"passRate":0.2127659574468085,"pass_rate":0.2127659574468085,"pass_rate_percent":21.27659574468085,"passes":30,"runs":134,"runtime":207594.6145255206,"source_flags":[],"source_harness":"droid","source_row_index":688,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":132,"tokens":null,"totalCostUsd":224.97,"totalDurationS":207594.6145255206,"totalInputTokens":214695571,"totalOutputTokens":2091602,"total_cost_usd":224.97,"total_input_tokens":214695571,"total_output_tokens":2091602,"total_runtime_s":207594.6145255206,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":41.36390008783094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-211358ec7ccfaeae4c9d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-90c0735c799326910ef19e53","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5353736949071799,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1278,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[638];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1278,"metadata":{"avgScore":0.5353736949071799,"avg_score":0.5353736949071799,"avg_score_percent":53.53736949071799,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":48.16,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":28164467,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":528351,"passRate":0.2537313432835821,"pass_rate":0.2537313432835821,"pass_rate_percent":25.37313432835821,"passes":17,"runs":67,"runtime":36974,"source_flags":[],"source_harness":"codex","source_row_index":638,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":48.16,"totalDurationS":36974,"totalInputTokens":28164467,"totalOutputTokens":528351,"total_cost_usd":48.16,"total_input_tokens":28164467,"total_output_tokens":528351,"total_runtime_s":36974,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.53736949071799,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-84e451d4ed5a38083b0b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-89489ae7d7bca8d13ad06dc0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.33301050066448096,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":190,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[94];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":190,"metadata":{"avgScore":0.33301050066448096,"avg_score":0.33301050066448096,"avg_score_percent":33.3010500664481,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.95,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":102418661,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1067638,"passRate":0.16363636363636364,"pass_rate":0.16363636363636364,"pass_rate_percent":16.363636363636363,"passes":9,"runs":54,"runtime":96800,"source_flags":[],"source_harness":"codex","source_row_index":94,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":127.95,"totalDurationS":96800,"totalInputTokens":102418661,"totalOutputTokens":1067638,"total_cost_usd":127.95,"total_input_tokens":102418661,"total_output_tokens":1067638,"total_runtime_s":96800,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":33.3010500664481,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e6e32bd288e00467e0b8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-847187b6c2d3cca8b0b2f6f3","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.15026470889226193,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1098,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[548];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1098,"metadata":{"avgScore":0.15026470889226193,"avg_score":0.15026470889226193,"avg_score_percent":15.026470889226193,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":231.75,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":262571368,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1892772.5,"passRate":0.016666666666666666,"pass_rate":0.016666666666666666,"pass_rate_percent":1.6666666666666667,"passes":1,"runs":59,"runtime":114277.5,"source_flags":[],"source_harness":"codex","source_row_index":548,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":231.75,"totalDurationS":114277.5,"totalInputTokens":262571368,"totalOutputTokens":1892772.5,"total_cost_usd":231.75,"total_input_tokens":262571368,"total_output_tokens":1892772.5,"total_runtime_s":114277.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":15.026470889226193,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e69399dfddc53476009","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7c1077a2645dc284afafce93","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3828945809408762,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1388,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[693];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1388,"metadata":{"avgScore":0.3828945809408762,"avg_score":0.3828945809408762,"avg_score_percent":38.28945809408762,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.9,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":96188305,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1176916,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":108014,"source_flags":[],"source_harness":"codex","source_row_index":693,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":127.9,"totalDurationS":108014,"totalInputTokens":96188305,"totalOutputTokens":1176916,"total_cost_usd":127.9,"total_input_tokens":96188305,"total_output_tokens":1176916,"total_runtime_s":108014,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.28945809408762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-51ce71b01629b715f3b9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-77d8d257fb7142d42a172d56","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.724007875487875,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":372,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[185];split=full/near-term;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":372,"metadata":{"avgScore":0.724007875487875,"avg_score":0.724007875487875,"avg_score_percent":72.4007875487875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":170.09,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":126210106.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2256939,"passRate":0.44029850746268656,"pass_rate":0.44029850746268656,"pass_rate_percent":44.02985074626866,"passes":58,"runs":132,"runtime":106336,"source_flags":[],"source_harness":"codex","source_row_index":185,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":170.09,"totalDurationS":106336,"totalInputTokens":126210106.5,"totalOutputTokens":2256939,"total_cost_usd":170.09,"total_input_tokens":126210106.5,"total_output_tokens":2256939,"total_runtime_s":106336,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":72.4007875487875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-311a6507980f9f6d2de2","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6e30d4de3230168f57e122e1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11836513333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1118,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[558];split=unlicensed/last-exam;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1118,"metadata":{"avgScore":0.11836513333333333,"avg_score":0.11836513333333333,"avg_score_percent":11.836513333333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.65,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":48704376,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":415252,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":25,"runtime":21599.263863500804,"source_flags":[],"source_harness":"cursor_cli","source_row_index":558,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":25,"tokens":null,"totalCostUsd":49.65,"totalDurationS":21599.263863500804,"totalInputTokens":48704376,"totalOutputTokens":415252,"total_cost_usd":49.65,"total_input_tokens":48704376,"total_output_tokens":415252,"total_runtime_s":21599.263863500804,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.836513333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5326a8c13dd7d9640dac","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6d622e162da2a7611b2ee2f6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.12697865037281922,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1112,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[555];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1112,"metadata":{"avgScore":0.12697865037281922,"avg_score":0.12697865037281922,"avg_score_percent":12.697865037281922,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.16,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":101477070,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":750493,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":29,"runtime":81523,"source_flags":[],"source_harness":"codex","source_row_index":555,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":110.16,"totalDurationS":81523,"totalInputTokens":101477070,"totalOutputTokens":750493,"total_cost_usd":110.16,"total_input_tokens":101477070,"total_output_tokens":750493,"total_runtime_s":81523,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":12.697865037281922,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9e61e6338e3815c4d498","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6d215d3cbcec9c4a7b2649a4","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.23112848158466556,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":614,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[306];split=licensed;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":614,"metadata":{"avgScore":0.23112848158466556,"avg_score":0.23112848158466556,"avg_score_percent":23.112848158466555,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":78.93,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":79134957.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":621202,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":21,"runtime":62259,"source_flags":[],"source_harness":"codex","source_row_index":306,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":78.93,"totalDurationS":62259,"totalInputTokens":79134957.5,"totalOutputTokens":621202,"total_cost_usd":78.93,"total_input_tokens":79134957.5,"total_output_tokens":621202,"total_runtime_s":62259,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.112848158466555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e9bd393905dcf0c7a7b3","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-685d2dfbd2c416895fd68fa5","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.44715784362305094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":42,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[20];split=full;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":42,"metadata":{"avgScore":0.44715784362305094,"avg_score":0.44715784362305094,"avg_score_percent":44.715784362305094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.59,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":489884949,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3517360.8333333335,"passRate":0.2445175438596491,"pass_rate":0.2445175438596491,"pass_rate_percent":24.45175438596491,"passes":110,"runs":438,"runtime":270900.40022482676,"source_flags":[],"source_harness":"codex","source_row_index":20,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":415.59,"totalDurationS":270900.40022482676,"totalInputTokens":489884949,"totalOutputTokens":3517360.8333333335,"total_cost_usd":415.59,"total_input_tokens":489884949,"total_output_tokens":3517360.8333333335,"total_runtime_s":270900.40022482676,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.715784362305094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2139fe43caa976765716","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-67395833f34c76d4d02825dc","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.480849060167781,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":61,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":740,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[369];split=linux_only;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":740,"metadata":{"avgScore":0.480849060167781,"avg_score":0.480849060167781,"avg_score_percent":48.0849060167781,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":108.61,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":77021220,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1313359,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":103,"runtime":60234.271691576985,"source_flags":[],"source_harness":"ale_claw","source_row_index":369,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":108.61,"totalDurationS":60234.271691576985,"totalInputTokens":77021220,"totalOutputTokens":1313359,"total_cost_usd":108.61,"total_input_tokens":77021220,"total_output_tokens":1313359,"total_runtime_s":60234.271691576985,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":48.0849060167781,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2860b111036327791180","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-60fd5883a1ecbb94badb0689","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.40481898174638836,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":62,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":784,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[391];split=linux_only;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":784,"metadata":{"avgScore":0.40481898174638836,"avg_score":0.40481898174638836,"avg_score_percent":40.481898174638836,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":135.91,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":117031110,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1468511,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":21,"runs":101,"runtime":134836.59417068138,"source_flags":[],"source_harness":"droid","source_row_index":391,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":135.91,"totalDurationS":134836.59417068138,"totalInputTokens":117031110,"totalOutputTokens":1468511,"total_cost_usd":135.91,"total_input_tokens":117031110,"total_output_tokens":1468511,"total_runtime_s":134836.59417068138,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":40.481898174638836,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bcccbb1c7351e0bbae1e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5bd3eaf432ceee898bec7943","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.4560896833299899,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":63,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":754,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[376];split=linux_only;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":754,"metadata":{"avgScore":0.4560896833299899,"avg_score":0.4560896833299899,"avg_score_percent":45.60896833299899,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":221.27,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":292732648,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2111892,"passRate":0.2365079365079365,"pass_rate":0.2365079365079365,"pass_rate_percent":23.65079365079365,"passes":74,"runs":310,"runtime":112697.86125335163,"source_flags":[],"source_harness":"codex","source_row_index":376,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":221.27,"totalDurationS":112697.86125335163,"totalInputTokens":292732648,"totalOutputTokens":2111892,"total_cost_usd":221.27,"total_input_tokens":292732648,"total_output_tokens":2111892,"total_runtime_s":112697.86125335163,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.60896833299899,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aa7c735e8f6618b002ba","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-59408b70dfdec344d586a525","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.49235206278939025,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":64,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":734,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[366];split=linux_only;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":734,"metadata":{"avgScore":0.49235206278939025,"avg_score":0.49235206278939025,"avg_score_percent":49.235206278939025,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":384.51,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":357080347,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":4337017,"passRate":0.2714285714285714,"pass_rate":0.2714285714285714,"pass_rate_percent":27.142857142857142,"passes":56,"runs":208,"runtime":171263.5,"source_flags":[],"source_harness":"codex","source_row_index":366,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":384.51,"totalDurationS":171263.5,"totalInputTokens":357080347,"totalOutputTokens":4337017,"total_cost_usd":384.51,"total_input_tokens":357080347,"total_output_tokens":4337017,"total_runtime_s":171263.5,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.235206278939025,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-32f251f0d950ce32bc0e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-53d2960aa70c69ffd58df20a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1459374165601684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":65,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1102,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[550];split=unlicensed/last-exam;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1102,"metadata":{"avgScore":0.1459374165601684,"avg_score":0.1459374165601684,"avg_score_percent":14.59374165601684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.38,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":91673091,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":597458,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":28,"runtime":41648.47447400001,"source_flags":[],"source_harness":"ale_claw","source_row_index":550,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":80.38,"totalDurationS":41648.47447400001,"totalInputTokens":91673091,"totalOutputTokens":597458,"total_cost_usd":80.38,"total_input_tokens":91673091,"total_output_tokens":597458,"total_runtime_s":41648.47447400001,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.59374165601684,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-14bbacda6e2c514f7983","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4fa188896103a74bbd885186","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3240193712304323,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":66,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":192,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[95];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":192,"metadata":{"avgScore":0.3240193712304323,"avg_score":0.3240193712304323,"avg_score_percent":32.401937123043226,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":45.35,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":29465113,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":425304,"passRate":0.16363636363636364,"pass_rate":0.16363636363636364,"pass_rate_percent":16.363636363636363,"passes":9,"runs":55,"runtime":53351,"source_flags":[],"source_harness":"codex","source_row_index":95,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":45.35,"totalDurationS":53351,"totalInputTokens":29465113,"totalOutputTokens":425304,"total_cost_usd":45.35,"total_input_tokens":29465113,"total_output_tokens":425304,"total_runtime_s":53351,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.401937123043226,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-eb1f0ecdff230c8121f1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4ad2b722e09839908163d27a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.42072396960513175,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":67,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":524,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[261];split=full/overall;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":524,"metadata":{"avgScore":0.42072396960513175,"avg_score":0.42072396960513175,"avg_score_percent":42.072396960513174,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446.52,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":469259907,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3279175,"passRate":0.21710526315789475,"pass_rate":0.21710526315789475,"pass_rate_percent":21.710526315789476,"passes":33,"runs":136,"runtime":297468.11566528457,"source_flags":[],"source_harness":"openclaw","source_row_index":261,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":136,"tokens":null,"totalCostUsd":446.52,"totalDurationS":297468.11566528457,"totalInputTokens":469259907,"totalOutputTokens":3279175,"total_cost_usd":446.52,"total_input_tokens":469259907,"total_output_tokens":3279175,"total_runtime_s":297468.11566528457,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.072396960513174,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-08a72caea805ef192c63","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-48e9dd4e236fe81537ef219c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.39352714009093803,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":68,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":992,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[495];split=unlicensed/full-spectrum;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":992,"metadata":{"avgScore":0.39352714009093803,"avg_score":0.39352714009093803,"avg_score_percent":39.3527140090938,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":53.27,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":37425146,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":628019,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":49,"runtime":24984.74481033199,"source_flags":[],"source_harness":"ale_claw","source_row_index":495,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":53.27,"totalDurationS":24984.74481033199,"totalInputTokens":37425146,"totalOutputTokens":628019,"total_cost_usd":53.27,"total_input_tokens":37425146,"total_output_tokens":628019,"total_runtime_s":24984.74481033199,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":39.3527140090938,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-66c7ee198db2a78dd266","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-46efa6cabb57523cf8b879a7","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3924651258147475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":69,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":544,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[271];split=full/overall;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":544,"metadata":{"avgScore":0.3924651258147475,"avg_score":0.3924651258147475,"avg_score_percent":39.246512581474754,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":242.24,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":234782616,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2196372,"passRate":0.19736842105263158,"pass_rate":0.19736842105263158,"pass_rate_percent":19.736842105263158,"passes":30,"runs":139,"runtime":229751.49587275658,"source_flags":[],"source_harness":"droid","source_row_index":271,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":242.24,"totalDurationS":229751.49587275658,"totalInputTokens":234782616,"totalOutputTokens":2196372,"total_cost_usd":242.24,"total_input_tokens":234782616,"total_output_tokens":2196372,"total_runtime_s":229751.49587275658,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":39.246512581474754,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9977ec77df57cc09e924","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4349c5562fb8394110ae4058","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.49867589404218615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":70,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":866,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[432];split=unlicensed;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":866,"metadata":{"avgScore":0.49867589404218615,"avg_score":0.49867589404218615,"avg_score_percent":49.86758940421861,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":522.76,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":481345798.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":5745106.5,"passRate":0.2872340425531915,"pass_rate":0.2872340425531915,"pass_rate_percent":28.723404255319153,"passes":80,"runs":278,"runtime":287393.5,"source_flags":[],"source_harness":"codex","source_row_index":432,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":522.76,"totalDurationS":287393.5,"totalInputTokens":481345798.5,"totalOutputTokens":5745106.5,"total_cost_usd":522.76,"total_input_tokens":481345798.5,"total_output_tokens":5745106.5,"total_runtime_s":287393.5,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.86758940421861,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9ac24c5a5b861195410e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-43107790d0dcb20cc0ca0caf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12848904115546353,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":71,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":654,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[326];split=licensed;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":654,"metadata":{"avgScore":0.12848904115546353,"avg_score":0.12848904115546353,"avg_score_percent":12.848904115546352,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":16.17,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":19445057,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":86437,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":6,"runtime":42157.90135035687,"source_flags":[],"source_harness":"cursor_cli","source_row_index":326,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":16.17,"totalDurationS":42157.90135035687,"totalInputTokens":19445057,"totalOutputTokens":86437,"total_cost_usd":16.17,"total_input_tokens":19445057,"total_output_tokens":86437,"total_runtime_s":42157.90135035687,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.848904115546352,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dcb4989af043a3c7a904","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3a110876fb958828ea90538c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12105454545454544,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":72,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":658,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[328];split=licensed;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":658,"metadata":{"avgScore":0.12105454545454544,"avg_score":0.12105454545454544,"avg_score_percent":12.105454545454544,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":17.28,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":20087045,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":104770,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":22156.88134723599,"source_flags":[],"source_harness":"droid","source_row_index":328,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":17.28,"totalDurationS":22156.88134723599,"totalInputTokens":20087045,"totalOutputTokens":104770,"total_cost_usd":17.28,"total_input_tokens":20087045,"total_output_tokens":104770,"total_runtime_s":22156.88134723599,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.105454545454544,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-283790e7aa6dba286183","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-37b6c0a32dc19a273165b913","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.39776664940409484,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":73,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":540,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[269];split=full/overall;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":540,"metadata":{"avgScore":0.39776664940409484,"avg_score":0.39776664940409484,"avg_score_percent":39.77666494040948,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":357.68,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":294404262,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2853487,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":149,"runtime":270641,"source_flags":[],"source_harness":"codex","source_row_index":269,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":357.68,"totalDurationS":270641,"totalInputTokens":294404262,"totalOutputTokens":2853487,"total_cost_usd":357.68,"total_input_tokens":294404262,"total_output_tokens":2853487,"total_runtime_s":270641,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.77666494040948,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0d08f2935339b36409ba","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-36e1fdb29f84c2a177e03c03","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.670526220421849,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":74,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":394,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[196];split=full/near-term;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":394,"metadata":{"avgScore":0.670526220421849,"avg_score":0.670526220421849,"avg_score_percent":67.0526220421849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":120.93,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":162729818,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1307423.8333333333,"passRate":0.3905472636815919,"pass_rate":0.3905472636815919,"pass_rate_percent":39.054726368159194,"passes":77,"runs":198,"runtime":97188.19844049853,"source_flags":[],"source_harness":"codex","source_row_index":196,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":120.93,"totalDurationS":97188.19844049853,"totalInputTokens":162729818,"totalOutputTokens":1307423.8333333333,"total_cost_usd":120.93,"total_input_tokens":162729818,"total_output_tokens":1307423.8333333333,"total_runtime_s":97188.19844049853,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.0526220421849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e62258a805b8e8a51c33","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-3491ee0833a58c79482ab351","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.3693544973841823,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":75,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1008,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[503];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1008,"metadata":{"avgScore":0.3693544973841823,"avg_score":0.3693544973841823,"avg_score_percent":36.93544973841823,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":90.3,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":91586748.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":987275.5,"passRate":0.22666666666666668,"pass_rate":0.22666666666666668,"pass_rate_percent":22.666666666666668,"passes":34,"runs":146,"runtime":46836.10881495468,"source_flags":[],"source_harness":"codex","source_row_index":503,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":90.3,"totalDurationS":46836.10881495468,"totalInputTokens":91586748.5,"totalOutputTokens":987275.5,"total_cost_usd":90.3,"total_input_tokens":91586748.5,"total_output_tokens":987275.5,"total_runtime_s":46836.10881495468,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.93544973841823,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-25e35ebba7945b440eef","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-326eda4ad21c6cbe04a307c9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.47658737865887824,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":76,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":872,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[435];split=unlicensed;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":872,"metadata":{"avgScore":0.47658737865887824,"avg_score":0.47658737865887824,"avg_score_percent":47.65873786588782,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":275.26,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":292991974,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2165400,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":138,"runtime":117204.74566453496,"source_flags":[],"source_harness":"ale_claw","source_row_index":435,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":275.26,"totalDurationS":117204.74566453496,"totalInputTokens":292991974,"totalOutputTokens":2165400,"total_cost_usd":275.26,"total_input_tokens":292991974,"total_output_tokens":2165400,"total_runtime_s":117204.74566453496,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":47.65873786588782,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-73a4172cc4a17962c0cc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-316ba157d064ff1aa53a1f5b","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.14192087907608547,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":77,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":268,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[133];split=full/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":268,"metadata":{"avgScore":0.14192087907608547,"avg_score":0.14192087907608547,"avg_score_percent":14.192087907608547,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":167.24,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":153886147,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":904996,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":37,"runtime":149774,"source_flags":[],"source_harness":"codex","source_row_index":133,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":167.24,"totalDurationS":149774,"totalInputTokens":153886147,"totalOutputTokens":904996,"total_cost_usd":167.24,"total_input_tokens":153886147,"total_output_tokens":904996,"total_runtime_s":149774,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.192087907608547,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2e98fdfddbf854bc9c49","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-304400ef8073190cd815bfa4","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.47931391024591824,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":78,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":30,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[14];split=full;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":30,"metadata":{"avgScore":0.47931391024591824,"avg_score":0.47931391024591824,"avg_score_percent":47.931391024591825,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":601.69,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":560480756,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":6366308.5,"passRate":0.26644736842105265,"pass_rate":0.26644736842105265,"pass_rate_percent":26.644736842105267,"passes":80,"runs":299,"runtime":349652.5,"source_flags":[],"source_harness":"codex","source_row_index":14,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":601.69,"totalDurationS":349652.5,"totalInputTokens":560480756,"totalOutputTokens":6366308.5,"total_cost_usd":601.69,"total_input_tokens":560480756,"total_output_tokens":6366308.5,"total_runtime_s":349652.5,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":47.931391024591825,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d8fdfb3ed02bc310c8c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2e84b87e59fd1291f06568c4","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.37550806805261017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":79,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":170,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[84];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":170,"metadata":{"avgScore":0.37550806805261017,"avg_score":0.37550806805261017,"avg_score_percent":37.55080680526102,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":130.35,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":130833923,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1174007,"passRate":0.22424242424242422,"pass_rate":0.22424242424242422,"pass_rate_percent":22.424242424242422,"passes":37,"runs":160,"runtime":75446.90317553736,"source_flags":[],"source_harness":"codex","source_row_index":84,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":130.35,"totalDurationS":75446.90317553736,"totalInputTokens":130833923,"totalOutputTokens":1174007,"total_cost_usd":130.35,"total_input_tokens":130833923,"total_output_tokens":1174007,"total_runtime_s":75446.90317553736,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.55080680526102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3a78536974b80b7b5932","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-289ee322f4a6093cfe963be8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.394105932064375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":80,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":788,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[393];split=linux_only;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":788,"metadata":{"avgScore":0.394105932064375,"avg_score":0.394105932064375,"avg_score_percent":39.4105932064375,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":96.38,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":75852337,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":901935,"passRate":0.18095238095238095,"pass_rate":0.18095238095238095,"pass_rate_percent":18.095238095238095,"passes":19,"runs":105,"runtime":59135,"source_flags":[],"source_harness":"codex","source_row_index":393,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":96.38,"totalDurationS":59135,"totalInputTokens":75852337,"totalOutputTokens":901935,"total_cost_usd":96.38,"total_input_tokens":75852337,"total_output_tokens":901935,"total_runtime_s":59135,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.4105932064375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a104ac7f1b14ad4f84cb","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-2791969c8cf7fd0e447bb903","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3636965114654929,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":81,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":552,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[275];split=full/overall;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":552,"metadata":{"avgScore":0.3636965114654929,"avg_score":0.3636965114654929,"avg_score_percent":36.369651146549295,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":146.61,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":109294354,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1270291,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":139211,"source_flags":[],"source_harness":"codex","source_row_index":275,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":146.61,"totalDurationS":139211,"totalInputTokens":109294354,"totalOutputTokens":1270291,"total_cost_usd":146.61,"total_input_tokens":109294354,"total_output_tokens":1270291,"total_runtime_s":139211,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.369651146549295,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fd39584447eaa7032f2a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-26286e2aee16f2f2a1e4d430","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.38707504844115914,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":82,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":164,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[81];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":164,"metadata":{"avgScore":0.38707504844115914,"avg_score":0.38707504844115914,"avg_score_percent":38.70750484411592,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":178.34,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":142776399,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2245802,"passRate":0.21818181818181817,"pass_rate":0.21818181818181817,"pass_rate_percent":21.818181818181817,"passes":24,"runs":107,"runtime":106548.5,"source_flags":[],"source_harness":"codex","source_row_index":81,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":178.34,"totalDurationS":106548.5,"totalInputTokens":142776399,"totalOutputTokens":2245802,"total_cost_usd":178.34,"total_input_tokens":142776399,"total_output_tokens":2245802,"total_runtime_s":106548.5,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.70750484411592,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a7de39be801abb4ff416","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-257d9f9b7892708bddb30d47","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.38708429229256636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":83,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1002,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[500];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1002,"metadata":{"avgScore":0.38708429229256636,"avg_score":0.38708429229256636,"avg_score_percent":38.70842922925664,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":135.4,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":100347290.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1853010,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":22,"runs":98,"runtime":74170.5,"source_flags":[],"source_harness":"codex","source_row_index":500,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":135.4,"totalDurationS":74170.5,"totalInputTokens":100347290.5,"totalOutputTokens":1853010,"total_cost_usd":135.4,"total_input_tokens":100347290.5,"total_output_tokens":1853010,"total_runtime_s":74170.5,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.70842922925664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-14daa06392d4b2addae9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2553ba97da2df15f4cefa6d8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10429084991739138,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":84,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1128,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[563];split=unlicensed/last-exam;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1128,"metadata":{"avgScore":0.10429084991739138,"avg_score":0.10429084991739138,"avg_score_percent":10.429084991739138,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.24,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":119273828,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":841764,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":25,"runtime":67603.12073842585,"source_flags":[],"source_harness":"openclaw","source_row_index":563,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":25,"tokens":null,"totalCostUsd":107.24,"totalDurationS":67603.12073842585,"totalInputTokens":119273828,"totalOutputTokens":841764,"total_cost_usd":107.24,"total_input_tokens":119273828,"total_output_tokens":841764,"total_runtime_s":67603.12073842585,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.429084991739138,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8afbce0b6f16cb406c74","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-22e775fbae3f14ea8143c736","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.6719239733700241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":85,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":392,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[195];split=full/near-term;harness=openclaw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":392,"metadata":{"avgScore":0.6719239733700241,"avg_score":0.6719239733700241,"avg_score_percent":67.1923973370024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":218.12,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":223015549,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1462808,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":148769.07770477302,"source_flags":[],"source_harness":"openclaw","source_row_index":195,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":218.12,"totalDurationS":148769.07770477302,"totalInputTokens":223015549,"totalOutputTokens":1462808,"total_cost_usd":218.12,"total_input_tokens":223015549,"total_output_tokens":1462808,"total_runtime_s":148769.07770477302,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":67.1923973370024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e984bb152c17f08963eb","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-225711d4cc34199ea9b25b38","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.11761216637194266,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":86,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":660,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[329];split=licensed;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":660,"metadata":{"avgScore":0.11761216637194266,"avg_score":0.11761216637194266,"avg_score_percent":11.761216637194266,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":18.71,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":13106049,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":93375,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":31197,"source_flags":[],"source_harness":"codex","source_row_index":329,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":18.71,"totalDurationS":31197,"totalInputTokens":13106049,"totalOutputTokens":93375,"total_cost_usd":18.71,"total_input_tokens":13106049,"total_output_tokens":93375,"total_runtime_s":31197,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.761216637194266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1b81020a12c8f81a877a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1f88fc986786115d372d7f23","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11267746971606471,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":87,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":284,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[141];split=full/last-exam;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":284,"metadata":{"avgScore":0.11267746971606471,"avg_score":0.11267746971606471,"avg_score_percent":11.26774697160647,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":75.62,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":83601379,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":551805,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":31,"runtime":95199.99296236214,"source_flags":[],"source_harness":"droid","source_row_index":141,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":29,"tokens":null,"totalCostUsd":75.62,"totalDurationS":95199.99296236214,"totalInputTokens":83601379,"totalOutputTokens":551805,"total_cost_usd":75.62,"total_input_tokens":83601379,"total_output_tokens":551805,"total_runtime_s":95199.99296236214,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.26774697160647,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c545d8dbccfdf4c0c996","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1ec1800b15c8b48608f97bb9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12633812830701532,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":88,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1114,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[556];split=unlicensed/last-exam;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1114,"metadata":{"avgScore":0.12633812830701532,"avg_score":0.12633812830701532,"avg_score_percent":12.633812830701533,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":65.27,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":71979435,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":506281,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":28,"runtime":75297.65644168308,"source_flags":[],"source_harness":"droid","source_row_index":556,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":65.27,"totalDurationS":75297.65644168308,"totalInputTokens":71979435,"totalOutputTokens":506281,"total_cost_usd":65.27,"total_input_tokens":71979435,"total_output_tokens":506281,"total_runtime_s":75297.65644168308,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.633812830701533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-42706c80333888ba36c8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1def2740b2343ccd2997b14a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.39445811616626686,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":89,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":786,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[392];split=linux_only;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":786,"metadata":{"avgScore":0.39445811616626686,"avg_score":0.39445811616626686,"avg_score_percent":39.445811616626685,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":194.65,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":157904128,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1928277,"passRate":0.17142857142857143,"pass_rate":0.17142857142857143,"pass_rate_percent":17.142857142857142,"passes":18,"runs":103,"runtime":95269,"source_flags":[],"source_harness":"codex","source_row_index":392,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":194.65,"totalDurationS":95269,"totalInputTokens":157904128,"totalOutputTokens":1928277,"total_cost_usd":194.65,"total_input_tokens":157904128,"total_output_tokens":1928277,"total_runtime_s":95269,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.445811616626685,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6c9d1620ed263d2b5b24","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-18c86b79af6037739db8c53d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5926809886791707,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":90,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":430,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[214];split=full/near-term;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":430,"metadata":{"avgScore":0.5926809886791707,"avg_score":0.5926809886791707,"avg_score_percent":59.268098867917075,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.43,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":81684093,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":1064184,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":66,"runtime":48118,"source_flags":[],"source_harness":"codex","source_row_index":214,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":110.43,"totalDurationS":48118,"totalInputTokens":81684093,"totalOutputTokens":1064184,"total_cost_usd":110.43,"total_input_tokens":81684093,"total_output_tokens":1064184,"total_runtime_s":48118,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":59.268098867917075,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-208a984bf0fc55a8cdbb","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0b2977efa8de5ca48b31b649","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4114434329247684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":91,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":908,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[453];split=unlicensed;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":908,"metadata":{"avgScore":0.4114434329247684,"avg_score":0.4114434329247684,"avg_score_percent":41.14434329247684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":291.95,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":235196044,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2640977,"passRate":0.19858156028368795,"pass_rate":0.19858156028368795,"pass_rate_percent":19.858156028368796,"passes":28,"runs":138,"runtime":179609,"source_flags":[],"source_harness":"codex","source_row_index":453,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":291.95,"totalDurationS":179609,"totalInputTokens":235196044,"totalOutputTokens":2640977,"total_cost_usd":291.95,"total_input_tokens":235196044,"total_output_tokens":2640977,"total_runtime_s":179609,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.14434329247684,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2a8504af275888be7b7a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-08e0acc6c3cc3b7ef21f5bbb","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.724007875487875,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":92,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1208,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[603];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1208,"metadata":{"avgScore":0.724007875487875,"avg_score":0.724007875487875,"avg_score_percent":72.4007875487875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":170.09,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":126210106.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2256939,"passRate":0.44029850746268656,"pass_rate":0.44029850746268656,"pass_rate_percent":44.02985074626866,"passes":58,"runs":132,"runtime":106336,"source_flags":[],"source_harness":"codex","source_row_index":603,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":170.09,"totalDurationS":106336,"totalInputTokens":126210106.5,"totalOutputTokens":2256939,"total_cost_usd":170.09,"total_input_tokens":126210106.5,"total_output_tokens":2256939,"total_runtime_s":106336,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":72.4007875487875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ae03278d83dcb959dc32","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-08643266fe930eee203f4498","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.31952009526446135,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":93,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1028,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[513];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1028,"metadata":{"avgScore":0.31952009526446135,"avg_score":0.31952009526446135,"avg_score_percent":31.952009526446133,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":82.36,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":58511596,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":965128,"passRate":0.16,"pass_rate":0.16,"pass_rate_percent":16.0,"passes":8,"runs":49,"runtime":55456,"source_flags":[],"source_harness":"codex","source_row_index":513,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":82.36,"totalDurationS":55456,"totalInputTokens":58511596,"totalOutputTokens":965128,"total_cost_usd":82.36,"total_input_tokens":58511596,"total_output_tokens":965128,"total_runtime_s":55456,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.952009526446133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0c5c9f83a1c31ddd1b02","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-07b50e24d631ba0b65ee92e9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4578443243363864,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":94,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":36,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[17];split=full;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":36,"metadata":{"avgScore":0.4578443243363864,"avg_score":0.4578443243363864,"avg_score_percent":45.78443243363864,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":309.77,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":334305043,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2340693,"passRate":0.23026315789473684,"pass_rate":0.23026315789473684,"pass_rate_percent":23.026315789473685,"passes":35,"runs":145,"runtime":128044.22844149194,"source_flags":[],"source_harness":"ale_claw","source_row_index":17,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":145,"tokens":null,"totalCostUsd":309.77,"totalDurationS":128044.22844149194,"totalInputTokens":334305043,"totalOutputTokens":2340693,"total_cost_usd":309.77,"total_input_tokens":334305043,"total_output_tokens":2340693,"total_runtime_s":128044.22844149194,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":45.78443243363864,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6fccfcaa12069704d67d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-06e86c03872b64266b75aa8f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4136390008783094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":95,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":906,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[452];split=unlicensed;harness=droid;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":906,"metadata":{"avgScore":0.4136390008783094,"avg_score":0.4136390008783094,"avg_score_percent":41.36390008783094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":224.97,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":214695571,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":2091602,"passRate":0.2127659574468085,"pass_rate":0.2127659574468085,"pass_rate_percent":21.27659574468085,"passes":30,"runs":134,"runtime":207594.6145255206,"source_flags":[],"source_harness":"droid","source_row_index":452,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":132,"tokens":null,"totalCostUsd":224.97,"totalDurationS":207594.6145255206,"totalInputTokens":214695571,"totalOutputTokens":2091602,"total_cost_usd":224.97,"total_input_tokens":214695571,"total_output_tokens":2091602,"total_runtime_s":207594.6145255206,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":41.36390008783094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7959f5900f662e9964c0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-02ef8d1b6421b2a04ad01321","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.46476081556733406,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a018930cb2c7f1007010c4b62c87e91d5838475","selectionRank":96,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1350,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[674];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1350,"metadata":{"avgScore":0.46476081556733406,"avg_score":0.46476081556733406,"avg_score_percent":46.476081556733405,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":339.05,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":417893865.5,"metric_id":"avg_score","metric_role":"secondary","model":"gpt-5-5","output_tokens":3171073.8333333335,"passRate":0.26122931442080377,"pass_rate":0.26122931442080377,"pass_rate_percent":26.12293144208038,"passes":109,"runs":412,"runtime":226290.44892996037,"source_flags":[],"source_harness":"codex","source_row_index":674,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":339.05,"totalDurationS":226290.44892996037,"totalInputTokens":417893865.5,"totalOutputTokens":3171073.8333333335,"total_cost_usd":339.05,"total_input_tokens":417893865.5,"total_output_tokens":3171073.8333333335,"total_runtime_s":226290.44892996037,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.476081556733405,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f32b113b5d043f01d159","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fce6996ee67e86c216eb2cd4","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23026315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":507,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[253];split=full/overall;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":507,"metadata":{"avgScore":0.4578443243363864,"avg_score":0.4578443243363864,"avg_score_percent":45.78443243363864,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":309.77,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":334305043,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2340693,"passRate":0.23026315789473684,"pass_rate":0.23026315789473684,"pass_rate_percent":23.026315789473685,"passes":35,"runs":145,"runtime":128044.22844149194,"source_flags":[],"source_harness":"ale_claw","source_row_index":253,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":145,"tokens":null,"totalCostUsd":309.77,"totalDurationS":128044.22844149194,"totalInputTokens":334305043,"totalOutputTokens":2340693,"total_cost_usd":309.77,"total_input_tokens":334305043,"total_output_tokens":2340693,"total_runtime_s":128044.22844149194,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.026315789473685,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c39ae3d903ab8bd9c8a1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fcdcd941ca8a1895414f6de8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1101,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[550];split=unlicensed/last-exam;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1101,"metadata":{"avgScore":0.1459374165601684,"avg_score":0.1459374165601684,"avg_score_percent":14.59374165601684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.38,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":91673091,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":597458,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":28,"runtime":41648.47447400001,"source_flags":[],"source_harness":"ale_claw","source_row_index":550,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":80.38,"totalDurationS":41648.47447400001,"totalInputTokens":91673091,"totalOutputTokens":597458,"total_cost_usd":80.38,"total_input_tokens":91673091,"total_output_tokens":597458,"total_runtime_s":41648.47447400001,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9636a3621cfc179dba45","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f41b2aaa506c77b828aa0b0a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2138157894736842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":531,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[265];split=full/overall;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":531,"metadata":{"avgScore":0.40680078557980925,"avg_score":0.40680078557980925,"avg_score_percent":40.68007855798093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":173.52,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":148061075,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1687764,"passRate":0.2138157894736842,"pass_rate":0.2138157894736842,"pass_rate_percent":21.38157894736842,"passes":33,"runs":142,"runtime":211627.30808669614,"source_flags":[],"source_harness":"cursor_cli","source_row_index":265,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":139,"tokens":null,"totalCostUsd":173.52,"totalDurationS":211627.30808669614,"totalInputTokens":148061075,"totalOutputTokens":1687764,"total_cost_usd":173.52,"total_input_tokens":148061075,"total_output_tokens":1687764,"total_runtime_s":211627.30808669614,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.38157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4f4fe3ad4fdc67a85230","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f4132774ab0e329028b5f945","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1001,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[500];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1001,"metadata":{"avgScore":0.38708429229256636,"avg_score":0.38708429229256636,"avg_score_percent":38.70842922925664,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":135.4,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":100347290.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1853010,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":22,"runs":98,"runtime":74170.5,"source_flags":[],"source_harness":"codex","source_row_index":500,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":135.4,"totalDurationS":74170.5,"totalInputTokens":100347290.5,"totalOutputTokens":1853010,"total_cost_usd":135.4,"total_input_tokens":100347290.5,"total_output_tokens":1853010,"total_runtime_s":74170.5,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cfde1e7cb9c14bb68c97","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f2732c8b866e47d0107d4e94","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24761904761904763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":747,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[373];split=linux_only;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":747,"metadata":{"avgScore":0.4605257773323447,"avg_score":0.4605257773323447,"avg_score_percent":46.052577733234465,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":218.69,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":211187431,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2152504,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":101,"runtime":113408.57987576415,"source_flags":[],"source_harness":"openclaw","source_row_index":373,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":218.69,"totalDurationS":113408.57987576415,"totalInputTokens":211187431,"totalOutputTokens":2152504,"total_cost_usd":218.69,"total_input_tokens":211187431,"total_output_tokens":2152504,"total_runtime_s":113408.57987576415,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.761904761904763,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0f06dfe8cd009ee28f25","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f223d4144bca6a8473f1b64f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":653,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[326];split=licensed;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":653,"metadata":{"avgScore":0.12848904115546353,"avg_score":0.12848904115546353,"avg_score_percent":12.848904115546352,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":16.17,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":19445057,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":86437,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":6,"runtime":42157.90135035687,"source_flags":[],"source_harness":"cursor_cli","source_row_index":326,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":6,"tokens":null,"totalCostUsd":16.17,"totalDurationS":42157.90135035687,"totalInputTokens":19445057,"totalOutputTokens":86437,"total_cost_usd":16.17,"total_input_tokens":19445057,"total_output_tokens":86437,"total_runtime_s":42157.90135035687,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0a595eb60a70badc585f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-f1d5290218f1132234715a69","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1139,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[569];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1139,"metadata":{"avgScore":0.09288872936875796,"avg_score":0.09288872936875796,"avg_score_percent":9.288872936875796,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":48.7,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":48327687,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":310243,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":30,"runtime":46146,"source_flags":[],"source_harness":"codex","source_row_index":569,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":48.7,"totalDurationS":46146,"totalInputTokens":48327687,"totalOutputTokens":310243,"total_cost_usd":48.7,"total_input_tokens":48327687,"total_output_tokens":310243,"total_runtime_s":46146,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-16aa4d6aa307da423330","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ee080f0ed3e7ca4bc0028053","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2445175438596491,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":41,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[20];split=full;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":41,"metadata":{"avgScore":0.44715784362305094,"avg_score":0.44715784362305094,"avg_score_percent":44.715784362305094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.59,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":489884949,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3517360.8333333335,"passRate":0.2445175438596491,"pass_rate":0.2445175438596491,"pass_rate_percent":24.45175438596491,"passes":110,"runs":438,"runtime":270900.40022482676,"source_flags":[],"source_harness":"codex","source_row_index":20,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":415.59,"totalDurationS":270900.40022482676,"totalInputTokens":489884949,"totalOutputTokens":3517360.8333333335,"total_cost_usd":415.59,"total_input_tokens":489884949,"total_output_tokens":3517360.8333333335,"total_runtime_s":270900.40022482676,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.45175438596491,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0d0d8a24cf108861aea0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ecd29bc62962ed49318884a5","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1015,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[507];split=unlicensed/full-spectrum;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1015,"metadata":{"avgScore":0.3503439333349274,"avg_score":0.3503439333349274,"avg_score_percent":35.034393333492744,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.39,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":49922490,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":723508,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":47,"runtime":53747.939585612694,"source_flags":[],"source_harness":"droid","source_row_index":507,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":47,"tokens":null,"totalCostUsd":61.39,"totalDurationS":53747.939585612694,"totalInputTokens":49922490,"totalOutputTokens":723508,"total_cost_usd":61.39,"total_input_tokens":49922490,"total_output_tokens":723508,"total_runtime_s":53747.939585612694,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b199bd407f64b5d39ce1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ecc351c9c0ca0e60bee1fb8f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.19858156028368795,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1379,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[689];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1379,"metadata":{"avgScore":0.4114434329247684,"avg_score":0.4114434329247684,"avg_score_percent":41.14434329247684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":291.95,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":235196044,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2640977,"passRate":0.19858156028368795,"pass_rate":0.19858156028368795,"pass_rate_percent":19.858156028368796,"passes":28,"runs":138,"runtime":179609,"source_flags":[],"source_harness":"codex","source_row_index":689,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":291.95,"totalDurationS":179609,"totalInputTokens":235196044,"totalOutputTokens":2640977,"total_cost_usd":291.95,"total_input_tokens":235196044,"total_output_tokens":2640977,"total_runtime_s":179609,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.858156028368796,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-40f0b2e672e2e22b384b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ec6ca73b3b0ba8d0a756aba3","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19736842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":71,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[35];split=full;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":71,"metadata":{"avgScore":0.3924651258147475,"avg_score":0.3924651258147475,"avg_score_percent":39.246512581474754,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":242.24,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":234782616,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2196372,"passRate":0.19736842105263158,"pass_rate":0.19736842105263158,"pass_rate_percent":19.736842105263158,"passes":30,"runs":139,"runtime":229751.49587275658,"source_flags":[],"source_harness":"droid","source_row_index":35,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":242.24,"totalDurationS":229751.49587275658,"totalInputTokens":234782616,"totalOutputTokens":2196372,"total_cost_usd":242.24,"total_input_tokens":234782616,"total_output_tokens":2196372,"total_runtime_s":229751.49587275658,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.736842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ca84d75407d6410acd1e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e8070661264628aa3f8a8b60","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1023,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[511];split=unlicensed/full-spectrum;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1023,"metadata":{"avgScore":0.3361494333333333,"avg_score":0.3361494333333333,"avg_score_percent":33.61494333333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":44.52,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":31812048,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":496590,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":10,"runs":48,"runtime":42562.361035950235,"source_flags":[],"source_harness":"cursor_cli","source_row_index":511,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":44.52,"totalDurationS":42562.361035950235,"totalInputTokens":31812048,"totalOutputTokens":496590,"total_cost_usd":44.52,"total_input_tokens":31812048,"total_output_tokens":496590,"total_runtime_s":42562.361035950235,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d22e997352ef9a765e87","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dfb2649b1aa4e04b6227ad31","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.373134328358209,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":391,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[195];split=full/near-term;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":391,"metadata":{"avgScore":0.6719239733700241,"avg_score":0.6719239733700241,"avg_score_percent":67.1923973370024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":218.12,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":223015549,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1462808,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":148769.07770477302,"source_flags":[],"source_harness":"openclaw","source_row_index":195,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":218.12,"totalDurationS":148769.07770477302,"totalInputTokens":223015549,"totalOutputTokens":1462808,"total_cost_usd":218.12,"total_input_tokens":223015549,"total_output_tokens":1462808,"total_runtime_s":148769.07770477302,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.3134328358209,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a7122a004906ed3bf445","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-df13e19b9a986580d7455ae8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18439716312056736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":915,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[457];split=unlicensed;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":915,"metadata":{"avgScore":0.3828945809408762,"avg_score":0.3828945809408762,"avg_score_percent":38.28945809408762,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.9,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":96188305,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1176916,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":108014,"source_flags":[],"source_harness":"codex","source_row_index":457,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":127.9,"totalDurationS":108014,"totalInputTokens":96188305,"totalOutputTokens":1176916,"total_cost_usd":127.9,"total_input_tokens":96188305,"total_output_tokens":1176916,"total_runtime_s":108014,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.439716312056735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1055039e152b487b274","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d54b74a2f88990f433939008","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.26122931442080377,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1349,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[674];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1349,"metadata":{"avgScore":0.46476081556733406,"avg_score":0.46476081556733406,"avg_score_percent":46.476081556733405,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":339.05,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":417893865.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3171073.8333333335,"passRate":0.26122931442080377,"pass_rate":0.26122931442080377,"pass_rate_percent":26.12293144208038,"passes":109,"runs":412,"runtime":226290.44892996037,"source_flags":[],"source_harness":"codex","source_row_index":674,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":339.05,"totalDurationS":226290.44892996037,"totalInputTokens":417893865.5,"totalOutputTokens":3171073.8333333335,"total_cost_usd":339.05,"total_input_tokens":417893865.5,"total_output_tokens":3171073.8333333335,"total_runtime_s":226290.44892996037,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.12293144208038,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-28144c005493a698ad23","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d1eb9bad409eb23f821c06fe","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1117,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[558];split=unlicensed/last-exam;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1117,"metadata":{"avgScore":0.11836513333333333,"avg_score":0.11836513333333333,"avg_score_percent":11.836513333333333,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.65,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":48704376,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":415252,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":25,"runtime":21599.263863500804,"source_flags":[],"source_harness":"cursor_cli","source_row_index":558,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":25,"tokens":null,"totalCostUsd":49.65,"totalDurationS":21599.263863500804,"totalInputTokens":48704376,"totalOutputTokens":415252,"total_cost_usd":49.65,"total_input_tokens":48704376,"total_output_tokens":415252,"total_runtime_s":21599.263863500804,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-269d8d9daecf711d3165","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-d110cb2776dc5bd242f59474","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2537313432835821,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":441,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[220];split=full/near-term;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":441,"metadata":{"avgScore":0.5353736949071799,"avg_score":0.5353736949071799,"avg_score_percent":53.53736949071799,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":48.16,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":28164467,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":528351,"passRate":0.2537313432835821,"pass_rate":0.2537313432835821,"pass_rate_percent":25.37313432835821,"passes":17,"runs":67,"runtime":36974,"source_flags":[],"source_harness":"codex","source_row_index":220,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":48.16,"totalDurationS":36974,"totalInputTokens":28164467,"totalOutputTokens":528351,"total_cost_usd":48.16,"total_input_tokens":28164467,"total_output_tokens":528351,"total_runtime_s":36974,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.37313432835821,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-493e774c9408c287b33a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-cf1deb7a8c426fd06bc0c4ff","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.17105263157894737,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":79,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[39];split=full;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":79,"metadata":{"avgScore":0.3636965114654929,"avg_score":0.3636965114654929,"avg_score_percent":36.369651146549295,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":146.61,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":109294354,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1270291,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":139211,"source_flags":[],"source_harness":"codex","source_row_index":39,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":146.61,"totalDurationS":139211,"totalInputTokens":109294354,"totalOutputTokens":1270291,"total_cost_usd":146.61,"total_input_tokens":109294354,"total_output_tokens":1270291,"total_runtime_s":139211,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.105263157894736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-935c5ced6d8a23fa2ec9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cbbee6012c9cc10b2e74b8e9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2127659574468085,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1377,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[688];split=unlicensed/overall;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1377,"metadata":{"avgScore":0.4136390008783094,"avg_score":0.4136390008783094,"avg_score_percent":41.36390008783094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":224.97,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":214695571,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2091602,"passRate":0.2127659574468085,"pass_rate":0.2127659574468085,"pass_rate_percent":21.27659574468085,"passes":30,"runs":134,"runtime":207594.6145255206,"source_flags":[],"source_harness":"droid","source_row_index":688,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":132,"tokens":null,"totalCostUsd":224.97,"totalDurationS":207594.6145255206,"totalInputTokens":214695571,"totalOutputTokens":2091602,"total_cost_usd":224.97,"total_input_tokens":214695571,"total_output_tokens":2091602,"total_runtime_s":207594.6145255206,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.27659574468085,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-66a735116dde582da410","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-ca97e36e7323b8d536821d1a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.013157894736842105,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":257,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[128];split=full/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":257,"metadata":{"avgScore":0.160932671435888,"avg_score":0.160932671435888,"avg_score_percent":16.0932671435888,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":283.24,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":316923871.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2234417.5,"passRate":0.013157894736842105,"pass_rate":0.013157894736842105,"pass_rate_percent":1.3157894736842104,"passes":1,"runs":74,"runtime":153534.5,"source_flags":[],"source_harness":"codex","source_row_index":128,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":283.24,"totalDurationS":153534.5,"totalInputTokens":316923871.5,"totalOutputTokens":2234417.5,"total_cost_usd":283.24,"total_input_tokens":316923871.5,"total_output_tokens":2234417.5,"total_runtime_s":153534.5,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":1.3157894736842104,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-169ef80bc1f0b689fa8b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c65ee792ac6e40bb242a8bab","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1127,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[563];split=unlicensed/last-exam;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1127,"metadata":{"avgScore":0.10429084991739138,"avg_score":0.10429084991739138,"avg_score_percent":10.429084991739138,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.24,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":119273828,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":841764,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":25,"runtime":67603.12073842585,"source_flags":[],"source_harness":"openclaw","source_row_index":563,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":25,"tokens":null,"totalCostUsd":107.24,"totalDurationS":67603.12073842585,"totalInputTokens":119273828,"totalOutputTokens":841764,"total_cost_usd":107.24,"total_input_tokens":119273828,"total_output_tokens":841764,"total_runtime_s":67603.12073842585,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a358bcb8eecf5133c9e1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c157fa305e90a49af4efe819","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":659,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[329];split=licensed;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":659,"metadata":{"avgScore":0.11761216637194266,"avg_score":0.11761216637194266,"avg_score_percent":11.761216637194266,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":18.71,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":13106049,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":93375,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":31197,"source_flags":[],"source_harness":"codex","source_row_index":329,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":18.71,"totalDurationS":31197,"totalInputTokens":13106049,"totalOutputTokens":93375,"total_cost_usd":18.71,"total_input_tokens":13106049,"total_output_tokens":93375,"total_runtime_s":31197,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4a83609fd908b0dbfdee","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c13bde3ef9b7c380f96ab7fe","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2365079365079365,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":753,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[376];split=linux_only;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":753,"metadata":{"avgScore":0.4560896833299899,"avg_score":0.4560896833299899,"avg_score_percent":45.60896833299899,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":221.27,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":292732648,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2111892,"passRate":0.2365079365079365,"pass_rate":0.2365079365079365,"pass_rate_percent":23.65079365079365,"passes":74,"runs":310,"runtime":112697.86125335163,"source_flags":[],"source_harness":"codex","source_row_index":376,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":221.27,"totalDurationS":112697.86125335163,"totalInputTokens":292732648,"totalOutputTokens":2111892,"total_cost_usd":221.27,"total_input_tokens":292732648,"total_output_tokens":2111892,"total_runtime_s":112697.86125335163,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.65079365079365,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-facbc08ecabc87682849","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c0fa1dbb7cd03af315811d48","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.44029850746268656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":371,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[185];split=full/near-term;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":371,"metadata":{"avgScore":0.724007875487875,"avg_score":0.724007875487875,"avg_score_percent":72.4007875487875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":170.09,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":126210106.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2256939,"passRate":0.44029850746268656,"pass_rate":0.44029850746268656,"pass_rate_percent":44.02985074626866,"passes":58,"runs":132,"runtime":106336,"source_flags":[],"source_harness":"codex","source_row_index":185,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":170.09,"totalDurationS":106336,"totalInputTokens":126210106.5,"totalOutputTokens":2256939,"total_cost_usd":170.09,"total_input_tokens":126210106.5,"total_output_tokens":2256939,"total_runtime_s":106336,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.02985074626866,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dc176a5f566746529497","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-bc9a4c8a2edc8995efef6370","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.21818181818181817,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":25,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":163,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[81];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":163,"metadata":{"avgScore":0.38707504844115914,"avg_score":0.38707504844115914,"avg_score_percent":38.70750484411592,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":178.34,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":142776399,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2245802,"passRate":0.21818181818181817,"pass_rate":0.21818181818181817,"pass_rate_percent":21.818181818181817,"passes":24,"runs":107,"runtime":106548.5,"source_flags":[],"source_harness":"codex","source_row_index":81,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":178.34,"totalDurationS":106548.5,"totalInputTokens":142776399,"totalOutputTokens":2245802,"total_cost_usd":178.34,"total_input_tokens":142776399,"total_output_tokens":2245802,"total_runtime_s":106548.5,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.818181818181817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d53771c48d1936c3eddc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bc00f187903bda5d025071ac","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":26,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1343,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[671];split=unlicensed/overall;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1343,"metadata":{"avgScore":0.47658737865887824,"avg_score":0.47658737865887824,"avg_score_percent":47.65873786588782,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":275.26,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":292991974,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2165400,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":138,"runtime":117204.74566453496,"source_flags":[],"source_harness":"ale_claw","source_row_index":671,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":275.26,"totalDurationS":117204.74566453496,"totalInputTokens":292991974,"totalOutputTokens":2165400,"total_cost_usd":275.26,"total_input_tokens":292991974,"total_output_tokens":2165400,"total_runtime_s":117204.74566453496,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9a8c79d108c36ca04b21","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b75c09c2d2970689e357e73d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":27,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":185,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[92];split=full/full-spectrum;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":185,"metadata":{"avgScore":0.33988274762503207,"avg_score":0.33988274762503207,"avg_score_percent":33.98827476250321,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":49.67,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":36415887,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":552100,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":11,"runs":52,"runtime":62650.35733736605,"source_flags":[],"source_harness":"cursor_cli","source_row_index":92,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":52,"tokens":null,"totalCostUsd":49.67,"totalDurationS":62650.35733736605,"totalInputTokens":36415887,"totalOutputTokens":552100,"total_cost_usd":49.67,"total_input_tokens":36415887,"total_output_tokens":552100,"total_runtime_s":62650.35733736605,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dce33a60248bcf1c9af5","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b1835335aaa28832c62d6218","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2127659574468085,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":28,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":905,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[452];split=unlicensed;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":905,"metadata":{"avgScore":0.4136390008783094,"avg_score":0.4136390008783094,"avg_score_percent":41.36390008783094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":224.97,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":214695571,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2091602,"passRate":0.2127659574468085,"pass_rate":0.2127659574468085,"pass_rate_percent":21.27659574468085,"passes":30,"runs":134,"runtime":207594.6145255206,"source_flags":[],"source_harness":"droid","source_row_index":452,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":132,"tokens":null,"totalCostUsd":224.97,"totalDurationS":207594.6145255206,"totalInputTokens":214695571,"totalOutputTokens":2091602,"total_cost_usd":224.97,"total_input_tokens":214695571,"total_output_tokens":2091602,"total_runtime_s":207594.6145255206,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.27659574468085,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-81de3ce0c0ac62609f18","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aaaf3e845d3325433b796dc0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.29850746268656714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":29,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1267,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[633];split=unlicensed/near-term;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1267,"metadata":{"avgScore":0.581507678464108,"avg_score":0.581507678464108,"avg_score_percent":58.1507678464108,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":105.98,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98383491,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":962554,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":65,"runtime":84340.52837546298,"source_flags":[],"source_harness":"droid","source_row_index":633,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":105.98,"totalDurationS":84340.52837546298,"totalInputTokens":98383491,"totalOutputTokens":962554,"total_cost_usd":105.98,"total_input_tokens":98383491,"total_output_tokens":962554,"total_runtime_s":84340.52837546298,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.850746268656714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-420e56281e6ee1829bf1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-aa5a3f0c7b09524a3711c3fa","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.016666666666666666,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":30,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1097,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[548];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1097,"metadata":{"avgScore":0.15026470889226193,"avg_score":0.15026470889226193,"avg_score_percent":15.026470889226193,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":231.75,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":262571368,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1892772.5,"passRate":0.016666666666666666,"pass_rate":0.016666666666666666,"pass_rate_percent":1.6666666666666667,"passes":1,"runs":59,"runtime":114277.5,"source_flags":[],"source_harness":"codex","source_row_index":548,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":231.75,"totalDurationS":114277.5,"totalInputTokens":262571368,"totalOutputTokens":1892772.5,"total_cost_usd":231.75,"total_input_tokens":262571368,"total_output_tokens":1892772.5,"total_runtime_s":114277.5,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":1.6666666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f8802985a0b8a0cead51","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a633dfbad563cf02812e6c14","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.26122931442080377,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":31,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":877,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[438];split=unlicensed;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":877,"metadata":{"avgScore":0.46476081556733406,"avg_score":0.46476081556733406,"avg_score_percent":46.476081556733405,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":339.05,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":417893865.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3171073.8333333335,"passRate":0.26122931442080377,"pass_rate":0.26122931442080377,"pass_rate_percent":26.12293144208038,"passes":109,"runs":412,"runtime":226290.44892996037,"source_flags":[],"source_harness":"codex","source_row_index":438,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":339.05,"totalDurationS":226290.44892996037,"totalInputTokens":417893865.5,"totalOutputTokens":3171073.8333333335,"total_cost_usd":339.05,"total_input_tokens":417893865.5,"total_output_tokens":3171073.8333333335,"total_runtime_s":226290.44892996037,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.12293144208038,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-008f24de2a24a24533ba","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-a5860f9c64380bcb01d17646","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2537313432835821,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":32,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1277,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[638];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1277,"metadata":{"avgScore":0.5353736949071799,"avg_score":0.5353736949071799,"avg_score_percent":53.53736949071799,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":48.16,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":28164467,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":528351,"passRate":0.2537313432835821,"pass_rate":0.2537313432835821,"pass_rate_percent":25.37313432835821,"passes":17,"runs":67,"runtime":36974,"source_flags":[],"source_harness":"codex","source_row_index":638,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":48.16,"totalDurationS":36974,"totalInputTokens":28164467,"totalOutputTokens":528351,"total_cost_usd":48.16,"total_input_tokens":28164467,"total_output_tokens":528351,"total_runtime_s":36974,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.37313432835821,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1747d8367349bd807e6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a38f6b8664f9035c51c0026f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23049645390070922,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":33,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":895,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[447];split=unlicensed;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":895,"metadata":{"avgScore":0.42851304932922624,"avg_score":0.42851304932922624,"avg_score_percent":42.85130493292262,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":157.35,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":128616018,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1601327,"passRate":0.23049645390070922,"pass_rate":0.23049645390070922,"pass_rate_percent":23.04964539007092,"passes":33,"runs":136,"runtime":169469.4067363392,"source_flags":[],"source_harness":"cursor_cli","source_row_index":447,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":133,"tokens":null,"totalCostUsd":157.35,"totalDurationS":169469.4067363392,"totalInputTokens":128616018,"totalOutputTokens":1601327,"total_cost_usd":157.35,"total_input_tokens":128616018,"total_output_tokens":1601327,"total_runtime_s":169469.4067363392,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.04964539007092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fd2189166b0c41ba16bc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-9d0fad82f3deabbaa0bb077d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18421052631578946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":34,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":539,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[269];split=full/overall;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":539,"metadata":{"avgScore":0.39776664940409484,"avg_score":0.39776664940409484,"avg_score_percent":39.77666494040948,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":357.68,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":294404262,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2853487,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":149,"runtime":270641,"source_flags":[],"source_harness":"codex","source_row_index":269,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":357.68,"totalDurationS":270641,"totalInputTokens":294404262,"totalOutputTokens":2853487,"total_cost_usd":357.68,"total_input_tokens":294404262,"total_output_tokens":2853487,"total_runtime_s":270641,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.421052631578945,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-68a621bd3e13ce21d72d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9cd60112dd96cb6fbc263bac","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":35,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":293,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[146];split=full/last-exam;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":293,"metadata":{"avgScore":0.10874277625057213,"avg_score":0.10874277625057213,"avg_score_percent":10.874277625057212,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":154.51,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":178227464,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":996954,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":29,"runtime":89528.58231328387,"source_flags":[],"source_harness":"openclaw","source_row_index":146,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":29,"tokens":null,"totalCostUsd":154.51,"totalDurationS":89528.58231328387,"totalInputTokens":178227464,"totalOutputTokens":996954,"total_cost_usd":154.51,"total_input_tokens":178227464,"total_output_tokens":996954,"total_runtime_s":89528.58231328387,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-45ab29594e76b85b9578","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9ad00d7aaed95e611fdaa292","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3283582089552239,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":36,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1225,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[612];split=unlicensed/near-term;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1225,"metadata":{"avgScore":0.6736635047623744,"avg_score":0.6736635047623744,"avg_score_percent":67.36635047623743,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":147.52,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":167219189,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1025815,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":67,"runtime":53801.491831077,"source_flags":[],"source_harness":"ale_claw","source_row_index":612,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":147.52,"totalDurationS":53801.491831077,"totalInputTokens":167219189,"totalOutputTokens":1025815,"total_cost_usd":147.52,"total_input_tokens":167219189,"total_output_tokens":1025815,"total_runtime_s":53801.491831077,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.83582089552239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fb216cdea586b44e952d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9807312bcf277c5c90b34e69","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":37,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":283,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[141];split=full/last-exam;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":283,"metadata":{"avgScore":0.11267746971606471,"avg_score":0.11267746971606471,"avg_score_percent":11.26774697160647,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":75.62,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":83601379,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":551805,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":31,"runtime":95199.99296236214,"source_flags":[],"source_harness":"droid","source_row_index":141,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":29,"tokens":null,"totalCostUsd":75.62,"totalDurationS":95199.99296236214,"totalInputTokens":83601379,"totalOutputTokens":551805,"total_cost_usd":75.62,"total_input_tokens":83601379,"total_output_tokens":551805,"total_runtime_s":95199.99296236214,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5dc22ad70e7c2648a185","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-943e66b37aa28ba384aaa870","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21710526315789475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":38,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":51,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[25];split=full;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":51,"metadata":{"avgScore":0.42072396960513175,"avg_score":0.42072396960513175,"avg_score_percent":42.072396960513174,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446.52,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":469259907,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3279175,"passRate":0.21710526315789475,"pass_rate":0.21710526315789475,"pass_rate_percent":21.710526315789476,"passes":33,"runs":136,"runtime":297468.11566528457,"source_flags":[],"source_harness":"openclaw","source_row_index":25,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":136,"tokens":null,"totalCostUsd":446.52,"totalDurationS":297468.11566528457,"totalInputTokens":469259907,"totalOutputTokens":3279175,"total_cost_usd":446.52,"total_input_tokens":469259907,"total_output_tokens":3279175,"total_runtime_s":297468.11566528457,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.710526315789476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e257c5a9bf8d8e5647d6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-93396e57f940b443ff037d31","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21710526315789475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":39,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":523,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[261];split=full/overall;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":523,"metadata":{"avgScore":0.42072396960513175,"avg_score":0.42072396960513175,"avg_score_percent":42.072396960513174,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":446.52,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":469259907,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3279175,"passRate":0.21710526315789475,"pass_rate":0.21710526315789475,"pass_rate_percent":21.710526315789476,"passes":33,"runs":136,"runtime":297468.11566528457,"source_flags":[],"source_harness":"openclaw","source_row_index":261,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":136,"tokens":null,"totalCostUsd":446.52,"totalDurationS":297468.11566528457,"totalInputTokens":469259907,"totalOutputTokens":3279175,"total_cost_usd":446.52,"total_input_tokens":469259907,"total_output_tokens":3279175,"total_runtime_s":297468.11566528457,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.710526315789476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2e89caf015d1f5d663ca","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8f5806c0311bfee761c0bd22","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.22666666666666668,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":40,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1007,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[503];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1007,"metadata":{"avgScore":0.3693544973841823,"avg_score":0.3693544973841823,"avg_score_percent":36.93544973841823,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":90.3,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":91586748.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":987275.5,"passRate":0.22666666666666668,"pass_rate":0.22666666666666668,"pass_rate_percent":22.666666666666668,"passes":34,"runs":146,"runtime":46836.10881495468,"source_flags":[],"source_harness":"codex","source_row_index":503,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":90.3,"totalDurationS":46836.10881495468,"totalInputTokens":91586748.5,"totalOutputTokens":987275.5,"total_cost_usd":90.3,"total_input_tokens":91586748.5,"total_output_tokens":987275.5,"total_runtime_s":46836.10881495468,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.666666666666668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0ffca19b2fd705a5b435","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8b780fce7a03ff3e8793fcc0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.29850746268656714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":41,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":431,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[215];split=full/near-term;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":431,"metadata":{"avgScore":0.581507678464108,"avg_score":0.581507678464108,"avg_score_percent":58.1507678464108,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":105.98,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98383491,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":962554,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":65,"runtime":84340.52837546298,"source_flags":[],"source_harness":"droid","source_row_index":215,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":105.98,"totalDurationS":84340.52837546298,"totalInputTokens":98383491,"totalOutputTokens":962554,"total_cost_usd":105.98,"total_input_tokens":98383491,"total_output_tokens":962554,"total_runtime_s":84340.52837546298,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":29.850746268656714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-195da71f09300a90a5e3","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-8a0b1d1ab96f540c070601c9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.22424242424242422,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":42,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":169,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[84];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":169,"metadata":{"avgScore":0.37550806805261017,"avg_score":0.37550806805261017,"avg_score_percent":37.55080680526102,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":130.35,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":130833923,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1174007,"passRate":0.22424242424242422,"pass_rate":0.22424242424242422,"pass_rate_percent":22.424242424242422,"passes":37,"runs":160,"runtime":75446.90317553736,"source_flags":[],"source_harness":"codex","source_row_index":84,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":130.35,"totalDurationS":75446.90317553736,"totalInputTokens":130833923,"totalOutputTokens":1174007,"total_cost_usd":130.35,"total_input_tokens":130833923,"total_output_tokens":1174007,"total_runtime_s":75446.90317553736,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.424242424242422,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6d5215dbc5396eeccd2a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-88a6dd7250399db905f0ad9f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":43,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":269,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[134];split=full/last-exam;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":269,"metadata":{"avgScore":0.13388631862928818,"avg_score":0.13388631862928818,"avg_score_percent":13.388631862928818,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":187.64,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":216406972,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1211914,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":103,"runtime":119659.35271642619,"source_flags":[],"source_harness":"codex","source_row_index":134,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":187.64,"totalDurationS":119659.35271642619,"totalInputTokens":216406972,"totalOutputTokens":1211914,"total_cost_usd":187.64,"total_input_tokens":216406972,"total_output_tokens":1211914,"total_runtime_s":119659.35271642619,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-af1cddb314d8e54097fa","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-844cc710f2801277fba3dba0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":44,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":613,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[306];split=licensed;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":613,"metadata":{"avgScore":0.23112848158466556,"avg_score":0.23112848158466556,"avg_score_percent":23.112848158466555,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":78.93,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":79134957.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":621202,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":21,"runtime":62259,"source_flags":[],"source_harness":"codex","source_row_index":306,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":78.93,"totalDurationS":62259,"totalInputTokens":79134957.5,"totalOutputTokens":621202,"total_cost_usd":78.93,"total_input_tokens":79134957.5,"total_output_tokens":621202,"total_runtime_s":62259,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1d2d57a78665995cd50","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7c308b545d3aab7b9fa47dfa","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18095238095238095,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":45,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":787,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[393];split=linux_only;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":787,"metadata":{"avgScore":0.394105932064375,"avg_score":0.394105932064375,"avg_score_percent":39.4105932064375,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":96.38,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":75852337,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":901935,"passRate":0.18095238095238095,"pass_rate":0.18095238095238095,"pass_rate_percent":18.095238095238095,"passes":19,"runs":105,"runtime":59135,"source_flags":[],"source_harness":"codex","source_row_index":393,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":96.38,"totalDurationS":59135,"totalInputTokens":75852337,"totalOutputTokens":901935,"total_cost_usd":96.38,"total_input_tokens":75852337,"total_output_tokens":901935,"total_runtime_s":59135,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.095238095238095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d22a2eee1f9925f5c58","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-7b75199eee3d3b244d3640d1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":46,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":267,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[133];split=full/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":267,"metadata":{"avgScore":0.14192087907608547,"avg_score":0.14192087907608547,"avg_score_percent":14.192087907608547,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":167.24,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":153886147,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":904996,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":37,"runtime":149774,"source_flags":[],"source_harness":"codex","source_row_index":133,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":167.24,"totalDurationS":149774,"totalInputTokens":153886147,"totalOutputTokens":904996,"total_cost_usd":167.24,"total_input_tokens":153886147,"total_output_tokens":904996,"total_runtime_s":149774,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3d94732d29376ac26025","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-78b769f6172159fa205e9160","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23404255319148937,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":47,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":885,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[442];split=unlicensed;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":885,"metadata":{"avgScore":0.44245775446794344,"avg_score":0.44245775446794344,"avg_score_percent":44.24577544679435,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":394.14,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":403857647,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3083904,"passRate":0.23404255319148937,"pass_rate":0.23404255319148937,"pass_rate_percent":23.404255319148938,"passes":33,"runs":131,"runtime":273966.4079745095,"source_flags":[],"source_harness":"openclaw","source_row_index":442,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":131,"tokens":null,"totalCostUsd":394.14,"totalDurationS":273966.4079745095,"totalInputTokens":403857647,"totalOutputTokens":3083904,"total_cost_usd":394.14,"total_input_tokens":403857647,"total_output_tokens":3083904,"total_runtime_s":273966.4079745095,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.404255319148938,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-92e6c823ecaef1ad5896","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7357f7d291d91f15bad01556","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.373134328358209,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":48,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1227,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[613];split=unlicensed/near-term;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1227,"metadata":{"avgScore":0.6719239733700241,"avg_score":0.6719239733700241,"avg_score_percent":67.1923973370024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":218.12,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":223015549,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1462808,"passRate":0.373134328358209,"pass_rate":0.373134328358209,"pass_rate_percent":37.3134328358209,"passes":25,"runs":65,"runtime":148769.07770477302,"source_flags":[],"source_harness":"openclaw","source_row_index":613,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":218.12,"totalDurationS":148769.07770477302,"totalInputTokens":223015549,"totalOutputTokens":1462808,"total_cost_usd":218.12,"total_input_tokens":223015549,"total_output_tokens":1462808,"total_runtime_s":148769.07770477302,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":37.3134328358209,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2a0506e08b4b727e4d56","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-72515ce73e869e22f2316772","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":49,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":295,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[147];split=full/last-exam;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":295,"metadata":{"avgScore":0.10732247368421054,"avg_score":0.10732247368421054,"avg_score_percent":10.732247368421053,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.49,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":63892750,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":456416,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":29,"runtime":61917.92565060669,"source_flags":[],"source_harness":"cursor_cli","source_row_index":147,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":29,"tokens":null,"totalCostUsd":61.49,"totalDurationS":61917.92565060669,"totalInputTokens":63892750,"totalOutputTokens":456416,"total_cost_usd":61.49,"total_input_tokens":63892750,"total_output_tokens":456416,"total_runtime_s":61917.92565060669,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c3de304d938b5fb74de","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6e84ef0f25059c5388926f44","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":50,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":649,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[324];split=licensed;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":649,"metadata":{"avgScore":0.14213636363636362,"avg_score":0.14213636363636362,"avg_score_percent":14.213636363636361,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":52.38,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":65402260,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":195271,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":23501.70769077503,"source_flags":[],"source_harness":"openclaw","source_row_index":324,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":52.38,"totalDurationS":23501.70769077503,"totalInputTokens":65402260,"totalOutputTokens":195271,"total_cost_usd":52.38,"total_input_tokens":65402260,"total_output_tokens":195271,"total_runtime_s":23501.70769077503,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-42cd51e7207f7cb25c12","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6bcb1e4a036a026d81576ce1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":51,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":657,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[328];split=licensed;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":657,"metadata":{"avgScore":0.12105454545454544,"avg_score":0.12105454545454544,"avg_score_percent":12.105454545454544,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":17.28,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":20087045,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":104770,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":5,"runtime":22156.88134723599,"source_flags":[],"source_harness":"droid","source_row_index":328,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":5,"tokens":null,"totalCostUsd":17.28,"totalDurationS":22156.88134723599,"totalInputTokens":20087045,"totalOutputTokens":104770,"total_cost_usd":17.28,"total_input_tokens":20087045,"total_output_tokens":104770,"total_runtime_s":22156.88134723599,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-29825b84b30950ef5af2","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-6b6c4c0b034765d560df4a1f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2872340425531915,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":52,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":865,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[432];split=unlicensed;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":865,"metadata":{"avgScore":0.49867589404218615,"avg_score":0.49867589404218615,"avg_score_percent":49.86758940421861,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":522.76,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":481345798.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":5745106.5,"passRate":0.2872340425531915,"pass_rate":0.2872340425531915,"pass_rate_percent":28.723404255319153,"passes":80,"runs":278,"runtime":287393.5,"source_flags":[],"source_harness":"codex","source_row_index":432,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":522.76,"totalDurationS":287393.5,"totalInputTokens":481345798.5,"totalOutputTokens":5745106.5,"total_cost_usd":522.76,"total_input_tokens":481345798.5,"total_output_tokens":5745106.5,"total_runtime_s":287393.5,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.723404255319153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-78e907764f1675ca030d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-67e1614bb36ff1348cbbdaf0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2714285714285714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":53,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":733,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[366];split=linux_only;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":733,"metadata":{"avgScore":0.49235206278939025,"avg_score":0.49235206278939025,"avg_score_percent":49.235206278939025,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":384.51,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":357080347,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":4337017,"passRate":0.2714285714285714,"pass_rate":0.2714285714285714,"pass_rate_percent":27.142857142857142,"passes":56,"runs":208,"runtime":171263.5,"source_flags":[],"source_harness":"codex","source_row_index":366,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":105,"tokens":null,"totalCostUsd":384.51,"totalDurationS":171263.5,"totalInputTokens":357080347,"totalOutputTokens":4337017,"total_cost_usd":384.51,"total_input_tokens":357080347,"total_output_tokens":4337017,"total_runtime_s":171263.5,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.142857142857142,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-592b390ae261f3cc1d07","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-67b305a59857aaef1c7564a2","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.26644736842105265,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":54,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":29,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[14];split=full;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":29,"metadata":{"avgScore":0.47931391024591824,"avg_score":0.47931391024591824,"avg_score_percent":47.931391024591825,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":601.69,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":560480756,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":6366308.5,"passRate":0.26644736842105265,"pass_rate":0.26644736842105265,"pass_rate_percent":26.644736842105267,"passes":80,"runs":299,"runtime":349652.5,"source_flags":[],"source_harness":"codex","source_row_index":14,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":601.69,"totalDurationS":349652.5,"totalInputTokens":560480756,"totalOutputTokens":6366308.5,"total_cost_usd":601.69,"total_input_tokens":560480756,"total_output_tokens":6366308.5,"total_runtime_s":349652.5,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.644736842105267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cde939658827f83fcbc2","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-65209da35e678cf12034290a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22857142857142856,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":55,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":765,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[382];split=linux_only;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":765,"metadata":{"avgScore":0.43427017991504974,"avg_score":0.43427017991504974,"avg_score_percent":43.427017991504975,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":97.48,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":68450446,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1079925,"passRate":0.22857142857142856,"pass_rate":0.22857142857142856,"pass_rate_percent":22.857142857142858,"passes":24,"runs":100,"runtime":52191.603746200475,"source_flags":[],"source_harness":"cursor_cli","source_row_index":382,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":100,"tokens":null,"totalCostUsd":97.48,"totalDurationS":52191.603746200475,"totalInputTokens":68450446,"totalOutputTokens":1079925,"total_cost_usd":97.48,"total_input_tokens":68450446,"total_output_tokens":1079925,"total_runtime_s":52191.603746200475,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.857142857142858,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6159c1f87bbcee6fc8de","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-61444b9c971a4a6e1d3bfa27","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":56,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":991,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[495];split=unlicensed/full-spectrum;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":991,"metadata":{"avgScore":0.39352714009093803,"avg_score":0.39352714009093803,"avg_score_percent":39.3527140090938,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":53.27,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":37425146,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":628019,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":49,"runtime":24984.74481033199,"source_flags":[],"source_harness":"ale_claw","source_row_index":495,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":53.27,"totalDurationS":24984.74481033199,"totalInputTokens":37425146,"totalOutputTokens":628019,"total_cost_usd":53.27,"total_input_tokens":37425146,"total_output_tokens":628019,"total_runtime_s":24984.74481033199,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5a7525567b625ee10dd7","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-60a9621be6d895a75fa5511a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.29850746268656714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":57,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":429,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[214];split=full/near-term;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":429,"metadata":{"avgScore":0.5926809886791707,"avg_score":0.5926809886791707,"avg_score_percent":59.268098867917075,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.43,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":81684093,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1064184,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":66,"runtime":48118,"source_flags":[],"source_harness":"codex","source_row_index":214,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":110.43,"totalDurationS":48118,"totalInputTokens":81684093,"totalOutputTokens":1064184,"total_cost_usd":110.43,"total_input_tokens":81684093,"total_output_tokens":1064184,"total_runtime_s":48118,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.850746268656714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0c8acf11eb8061a7f55c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5ff015662c4b2357dc5e2114","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2872340425531915,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":58,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1337,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[668];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1337,"metadata":{"avgScore":0.49867589404218615,"avg_score":0.49867589404218615,"avg_score_percent":49.86758940421861,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":522.76,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":481345798.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":5745106.5,"passRate":0.2872340425531915,"pass_rate":0.2872340425531915,"pass_rate_percent":28.723404255319153,"passes":80,"runs":278,"runtime":287393.5,"source_flags":[],"source_harness":"codex","source_row_index":668,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":522.76,"totalDurationS":287393.5,"totalInputTokens":481345798.5,"totalOutputTokens":5745106.5,"total_cost_usd":522.76,"total_input_tokens":481345798.5,"total_output_tokens":5745106.5,"total_runtime_s":287393.5,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.723404255319153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fda23eab5c5ec0e58412","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-5bd60d8acc4afc95b47fd80d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.44029850746268656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":59,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1207,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[603];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1207,"metadata":{"avgScore":0.724007875487875,"avg_score":0.724007875487875,"avg_score_percent":72.4007875487875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":170.09,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":126210106.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2256939,"passRate":0.44029850746268656,"pass_rate":0.44029850746268656,"pass_rate_percent":44.02985074626866,"passes":58,"runs":132,"runtime":106336,"source_flags":[],"source_harness":"codex","source_row_index":603,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":170.09,"totalDurationS":106336,"totalInputTokens":126210106.5,"totalOutputTokens":2256939,"total_cost_usd":170.09,"total_input_tokens":126210106.5,"total_output_tokens":2256939,"total_runtime_s":106336,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.02985074626866,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-98f2c92ae42793c140a8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-576f46906e359206750cc663","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23404255319148937,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":60,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1357,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[678];split=unlicensed/overall;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1357,"metadata":{"avgScore":0.44245775446794344,"avg_score":0.44245775446794344,"avg_score_percent":44.24577544679435,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":394.14,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":403857647,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3083904,"passRate":0.23404255319148937,"pass_rate":0.23404255319148937,"pass_rate_percent":23.404255319148938,"passes":33,"runs":131,"runtime":273966.4079745095,"source_flags":[],"source_harness":"openclaw","source_row_index":678,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":131,"tokens":null,"totalCostUsd":394.14,"totalDurationS":273966.4079745095,"totalInputTokens":403857647,"totalOutputTokens":3083904,"total_cost_usd":394.14,"total_input_tokens":403857647,"total_output_tokens":3083904,"total_runtime_s":273966.4079745095,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.404255319148938,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fd57d2e3dd90f2533dfd","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-50af39985d7f11e94978adb6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.29850746268656714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":61,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1265,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[632];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1265,"metadata":{"avgScore":0.5926809886791707,"avg_score":0.5926809886791707,"avg_score_percent":59.268098867917075,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.43,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":81684093,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1064184,"passRate":0.29850746268656714,"pass_rate":0.29850746268656714,"pass_rate_percent":29.850746268656714,"passes":20,"runs":66,"runtime":48118,"source_flags":[],"source_harness":"codex","source_row_index":632,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":110.43,"totalDurationS":48118,"totalInputTokens":81684093,"totalOutputTokens":1064184,"total_cost_usd":110.43,"total_input_tokens":81684093,"total_output_tokens":1064184,"total_runtime_s":48118,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.850746268656714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-345a0ec3b8c5110e7f67","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-491772ed2974f5ba39af7baf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":62,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":783,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[391];split=linux_only;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":783,"metadata":{"avgScore":0.40481898174638836,"avg_score":0.40481898174638836,"avg_score_percent":40.481898174638836,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":135.91,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":117031110,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1468511,"passRate":0.2,"pass_rate":0.2,"pass_rate_percent":20.0,"passes":21,"runs":101,"runtime":134836.59417068138,"source_flags":[],"source_harness":"droid","source_row_index":391,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":101,"tokens":null,"totalCostUsd":135.91,"totalDurationS":134836.59417068138,"totalInputTokens":117031110,"totalOutputTokens":1468511,"total_cost_usd":135.91,"total_input_tokens":117031110,"total_output_tokens":1468511,"total_runtime_s":134836.59417068138,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c967a9e3fdb1366730f8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-48eaab608fdfedd7a7589068","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":63,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":615,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[307];split=licensed;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":615,"metadata":{"avgScore":0.22245515154818696,"avg_score":0.22245515154818696,"avg_score_percent":22.245515154818698,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":65.73,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":59208218,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":212510,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":11,"runtime":91032,"source_flags":[],"source_harness":"codex","source_row_index":307,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":65.73,"totalDurationS":91032,"totalInputTokens":59208218,"totalOutputTokens":212510,"total_cost_usd":65.73,"total_input_tokens":59208218,"total_output_tokens":212510,"total_runtime_s":91032,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1e16e1b0af06a6b25987","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-48baba81bf7b892d0fc5bfd9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.3905472636815919,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":64,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1229,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[614];split=unlicensed/near-term;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1229,"metadata":{"avgScore":0.670526220421849,"avg_score":0.670526220421849,"avg_score_percent":67.0526220421849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":120.93,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":162729818,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1307423.8333333333,"passRate":0.3905472636815919,"pass_rate":0.3905472636815919,"pass_rate_percent":39.054726368159194,"passes":77,"runs":198,"runtime":97188.19844049853,"source_flags":[],"source_harness":"codex","source_row_index":614,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":120.93,"totalDurationS":97188.19844049853,"totalInputTokens":162729818,"totalOutputTokens":1307423.8333333333,"total_cost_usd":120.93,"total_input_tokens":162729818,"total_output_tokens":1307423.8333333333,"total_runtime_s":97188.19844049853,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.054726368159194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-de958eb4c372793c4167","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-47c8f891b40d11a2892386d8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23026315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":65,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":35,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[17];split=full;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":35,"metadata":{"avgScore":0.4578443243363864,"avg_score":0.4578443243363864,"avg_score_percent":45.78443243363864,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":309.77,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":334305043,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2340693,"passRate":0.23026315789473684,"pass_rate":0.23026315789473684,"pass_rate_percent":23.026315789473685,"passes":35,"runs":145,"runtime":128044.22844149194,"source_flags":[],"source_harness":"ale_claw","source_row_index":17,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":145,"tokens":null,"totalCostUsd":309.77,"totalDurationS":128044.22844149194,"totalInputTokens":334305043,"totalOutputTokens":2340693,"total_cost_usd":309.77,"total_input_tokens":334305043,"total_output_tokens":2340693,"total_runtime_s":128044.22844149194,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.026315789473685,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d17e1a61f3add507a11a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-45273783fc2689668b7c3ffa","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24113475177304963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":66,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":871,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[435];split=unlicensed;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":871,"metadata":{"avgScore":0.47658737865887824,"avg_score":0.47658737865887824,"avg_score_percent":47.65873786588782,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":275.26,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":292991974,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2165400,"passRate":0.24113475177304963,"pass_rate":0.24113475177304963,"pass_rate_percent":24.113475177304963,"passes":34,"runs":138,"runtime":117204.74566453496,"source_flags":[],"source_harness":"ale_claw","source_row_index":435,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":275.26,"totalDurationS":117204.74566453496,"totalInputTokens":292991974,"totalOutputTokens":2165400,"total_cost_usd":275.26,"total_input_tokens":292991974,"total_output_tokens":2165400,"total_runtime_s":117204.74566453496,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.113475177304963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bcc667654043d92d399e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-43b67927245ae58c5d3d1b39","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23049645390070922,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":67,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1367,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[683];split=unlicensed/overall;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1367,"metadata":{"avgScore":0.42851304932922624,"avg_score":0.42851304932922624,"avg_score_percent":42.85130493292262,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":157.35,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":128616018,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1601327,"passRate":0.23049645390070922,"pass_rate":0.23049645390070922,"pass_rate_percent":23.04964539007092,"passes":33,"runs":136,"runtime":169469.4067363392,"source_flags":[],"source_harness":"cursor_cli","source_row_index":683,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":133,"tokens":null,"totalCostUsd":157.35,"totalDurationS":169469.4067363392,"totalInputTokens":128616018,"totalOutputTokens":1601327,"total_cost_usd":157.35,"total_input_tokens":128616018,"total_output_tokens":1601327,"total_runtime_s":169469.4067363392,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.04964539007092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d12c4ec9fd4de7556a7f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-408d40b3f198dc7f17e5f9fd","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":68,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1019,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[509];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1019,"metadata":{"avgScore":0.34037980623548364,"avg_score":0.34037980623548364,"avg_score_percent":34.037980623548364,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":35.01,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":21829853,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":377771,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":50,"runtime":28603,"source_flags":[],"source_harness":"codex","source_row_index":509,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":35.01,"totalDurationS":28603,"totalInputTokens":21829853,"totalOutputTokens":377771,"total_cost_usd":35.01,"total_input_tokens":21829853,"total_output_tokens":377771,"total_runtime_s":28603,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-18edcc7f8dfc9f200d19","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3f93fe5510cbdee4a826f88f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3358208955223881,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":69,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1245,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[622];split=unlicensed/near-term;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1245,"metadata":{"avgScore":0.6228643923694662,"avg_score":0.6228643923694662,"avg_score_percent":62.28643923694662,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.46,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":51827010,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":741097,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":23,"runs":69,"runtime":108587.11174137115,"source_flags":[],"source_harness":"cursor_cli","source_row_index":622,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":68.46,"totalDurationS":108587.11174137115,"totalInputTokens":51827010,"totalOutputTokens":741097,"total_cost_usd":68.46,"total_input_tokens":51827010,"total_output_tokens":741097,"total_runtime_s":108587.11174137115,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.582089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ef6add06172b7c8e57d9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-398807bbfe33100598f10f8d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18439716312056736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":70,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1387,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[693];split=unlicensed/overall;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1387,"metadata":{"avgScore":0.3828945809408762,"avg_score":0.3828945809408762,"avg_score_percent":38.28945809408762,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.9,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":96188305,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1176916,"passRate":0.18439716312056736,"pass_rate":0.18439716312056736,"pass_rate_percent":18.439716312056735,"passes":26,"runs":141,"runtime":108014,"source_flags":[],"source_harness":"codex","source_row_index":693,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":141,"tokens":null,"totalCostUsd":127.9,"totalDurationS":108014,"totalInputTokens":96188305,"totalOutputTokens":1176916,"total_cost_usd":127.9,"total_input_tokens":96188305,"total_output_tokens":1176916,"total_runtime_s":108014,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.439716312056735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-694fa13543387a046174","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-390f188828eba0f9315e527c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":71,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":621,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[310];split=licensed;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":621,"metadata":{"avgScore":0.21759244620262602,"avg_score":0.21759244620262602,"avg_score_percent":21.759244620262603,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":34.5,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":41313069,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":175293,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":7,"runtime":10839.482776957002,"source_flags":[],"source_harness":"ale_claw","source_row_index":310,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":34.5,"totalDurationS":10839.482776957002,"totalInputTokens":41313069,"totalOutputTokens":175293,"total_cost_usd":34.5,"total_input_tokens":41313069,"total_output_tokens":175293,"total_runtime_s":10839.482776957002,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c23053b3ac48397f982a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-34cad9e099d658ad716fb75c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.17105263157894737,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":72,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":551,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[275];split=full/overall;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":551,"metadata":{"avgScore":0.3636965114654929,"avg_score":0.3636965114654929,"avg_score_percent":36.369651146549295,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":146.61,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":109294354,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1270291,"passRate":0.17105263157894737,"pass_rate":0.17105263157894737,"pass_rate_percent":17.105263157894736,"passes":26,"runs":152,"runtime":139211,"source_flags":[],"source_harness":"codex","source_row_index":275,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":146.61,"totalDurationS":139211,"totalInputTokens":109294354,"totalOutputTokens":1270291,"total_cost_usd":146.61,"total_input_tokens":109294354,"total_output_tokens":1270291,"total_runtime_s":139211,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.105263157894736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e05c72d5db7a331e421f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-347f7ff89b2c571d6e2cd573","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":73,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":305,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[152];split=full/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":305,"metadata":{"avgScore":0.09880317382248731,"avg_score":0.09880317382248731,"avg_score_percent":9.88031738224873,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":64.34,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":59361487,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":378037,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":38,"runtime":57793,"source_flags":[],"source_harness":"codex","source_row_index":152,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":38,"tokens":null,"totalCostUsd":64.34,"totalDurationS":57793,"totalInputTokens":59361487,"totalOutputTokens":378037,"total_cost_usd":64.34,"total_input_tokens":59361487,"total_output_tokens":378037,"total_runtime_s":57793,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-69e8d426aefce5eeec0a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-32d574b7bf1975691dbc1208","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2138157894736842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":74,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":59,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[29];split=full;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":59,"metadata":{"avgScore":0.40680078557980925,"avg_score":0.40680078557980925,"avg_score_percent":40.68007855798093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":173.52,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":148061075,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1687764,"passRate":0.2138157894736842,"pass_rate":0.2138157894736842,"pass_rate_percent":21.38157894736842,"passes":33,"runs":142,"runtime":211627.30808669614,"source_flags":[],"source_harness":"cursor_cli","source_row_index":29,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":139,"tokens":null,"totalCostUsd":173.52,"totalDurationS":211627.30808669614,"totalInputTokens":148061075,"totalOutputTokens":1687764,"total_cost_usd":173.52,"total_input_tokens":148061075,"total_output_tokens":1687764,"total_runtime_s":211627.30808669614,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.38157894736842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fc1cc7b9d0f12cfa1ff0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-31793bb614cc5fec4e8cdc70","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.16,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":75,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1027,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[513];split=unlicensed/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1027,"metadata":{"avgScore":0.31952009526446135,"avg_score":0.31952009526446135,"avg_score_percent":31.952009526446133,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":82.36,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":58511596,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":965128,"passRate":0.16,"pass_rate":0.16,"pass_rate_percent":16.0,"passes":8,"runs":49,"runtime":55456,"source_flags":[],"source_harness":"codex","source_row_index":513,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":82.36,"totalDurationS":55456,"totalInputTokens":58511596,"totalOutputTokens":965128,"total_cost_usd":82.36,"total_input_tokens":58511596,"total_output_tokens":965128,"total_runtime_s":55456,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ca9d811e408447db0569","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2add8dfd12b6ce46dad41891","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"cursor_cli","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3358208955223881,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":76,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":409,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[204];split=full/near-term;harness=cursor_cli;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":409,"metadata":{"avgScore":0.6228643923694662,"avg_score":0.6228643923694662,"avg_score_percent":62.28643923694662,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.46,"costSource":"list","cost_source":"list","harness":"cursor_cli","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":51827010,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":741097,"passRate":0.3358208955223881,"pass_rate":0.3358208955223881,"pass_rate_percent":33.582089552238806,"passes":23,"runs":69,"runtime":108587.11174137115,"source_flags":[],"source_harness":"cursor_cli","source_row_index":204,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":68.46,"totalDurationS":108587.11174137115,"totalInputTokens":51827010,"totalOutputTokens":741097,"total_cost_usd":68.46,"total_input_tokens":51827010,"total_output_tokens":741097,"total_runtime_s":108587.11174137115,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.582089552238806,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-46600586308c9e4cc9be","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2302f983e9feb22906dec84c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":77,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":273,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[136];split=full/last-exam;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":273,"metadata":{"avgScore":0.12815059202118556,"avg_score":0.12815059202118556,"avg_score_percent":12.815059202118556,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":107.32,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":126287850,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":704360,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":32,"runtime":49458.08196799901,"source_flags":[],"source_harness":"ale_claw","source_row_index":136,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":32,"tokens":null,"totalCostUsd":107.32,"totalDurationS":49458.08196799901,"totalInputTokens":126287850,"totalOutputTokens":704360,"total_cost_usd":107.32,"total_input_tokens":126287850,"total_output_tokens":704360,"total_runtime_s":49458.08196799901,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-73983ba4f60e716ae80f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1f9143aff81ef0c48bd00089","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.2445175438596491,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":78,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":513,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[256];split=full/overall;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":513,"metadata":{"avgScore":0.44715784362305094,"avg_score":0.44715784362305094,"avg_score_percent":44.715784362305094,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":415.59,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":489884949,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":3517360.8333333335,"passRate":0.2445175438596491,"pass_rate":0.2445175438596491,"pass_rate_percent":24.45175438596491,"passes":110,"runs":438,"runtime":270900.40022482676,"source_flags":[],"source_harness":"codex","source_row_index":256,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":415.59,"totalDurationS":270900.40022482676,"totalInputTokens":489884949,"totalOutputTokens":3517360.8333333335,"total_cost_usd":415.59,"total_input_tokens":489884949,"total_output_tokens":3517360.8333333335,"total_runtime_s":270900.40022482676,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.45175438596491,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-df5b62be61466890cf42","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1e4dc1a6f37e7c1211382fdf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":79,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1113,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[556];split=unlicensed/last-exam;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1113,"metadata":{"avgScore":0.12633812830701532,"avg_score":0.12633812830701532,"avg_score_percent":12.633812830701533,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":65.27,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":71979435,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":506281,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":28,"runtime":75297.65644168308,"source_flags":[],"source_harness":"droid","source_row_index":556,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":26,"tokens":null,"totalCostUsd":65.27,"totalDurationS":75297.65644168308,"totalInputTokens":71979435,"totalOutputTokens":506281,"total_cost_usd":65.27,"total_input_tokens":71979435,"total_output_tokens":506281,"total_runtime_s":75297.65644168308,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e6edbab4590a9f49465b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1dcec5d8aa2972aa1f28a335","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":80,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":187,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[93];split=full/full-spectrum;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":187,"metadata":{"avgScore":0.333952530257971,"avg_score":0.333952530257971,"avg_score_percent":33.3952530257971,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":101.27,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98557834,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1053767,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":50,"runtime":82657.90435112063,"source_flags":[],"source_harness":"openclaw","source_row_index":93,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":101.27,"totalDurationS":82657.90435112063,"totalInputTokens":98557834,"totalOutputTokens":1053767,"total_cost_usd":101.27,"total_input_tokens":98557834,"total_output_tokens":1053767,"total_runtime_s":82657.90435112063,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-78eeab035af41ccbc49b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1d8cc0e3bdc8fa388a6bdf6c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.0303030303030303,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":81,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":619,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[309];split=licensed;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":619,"metadata":{"avgScore":0.22151974870087496,"avg_score":0.22151974870087496,"avg_score_percent":22.151974870087496,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":76.54,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":71991083.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":346287,"passRate":0.0303030303030303,"pass_rate":0.0303030303030303,"pass_rate_percent":3.03030303030303,"passes":1,"runs":26,"runtime":44609.951294866376,"source_flags":[],"source_harness":"codex","source_row_index":309,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":76.54,"totalDurationS":44609.951294866376,"totalInputTokens":71991083.5,"totalOutputTokens":346287,"total_cost_usd":76.54,"total_input_tokens":71991083.5,"total_output_tokens":346287,"total_runtime_s":44609.951294866376,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.03030303030303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8422ac0ccb6217cad93","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1c024bb13377496ef8234973","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.19858156028368795,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":82,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":907,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[453];split=unlicensed;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":907,"metadata":{"avgScore":0.4114434329247684,"avg_score":0.4114434329247684,"avg_score_percent":41.14434329247684,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":291.95,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":235196044,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2640977,"passRate":0.19858156028368795,"pass_rate":0.19858156028368795,"pass_rate_percent":19.858156028368796,"passes":28,"runs":138,"runtime":179609,"source_flags":[],"source_harness":"codex","source_row_index":453,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":291.95,"totalDurationS":179609,"totalInputTokens":235196044,"totalOutputTokens":2640977,"total_cost_usd":291.95,"total_input_tokens":235196044,"total_output_tokens":2640977,"total_runtime_s":179609,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.858156028368796,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3cf4c747ede7475abebe","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-17cad275de588ae8de3564cc","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-xhigh","source_harness":"codex","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.26644736842105265,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":83,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":501,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[250];split=full/overall;harness=codex;model=gpt-5-5;variant=reasoning-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":501,"metadata":{"avgScore":0.47931391024591824,"avg_score":0.47931391024591824,"avg_score_percent":47.931391024591825,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":601.69,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-xhigh","input_tokens":560480756,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":6366308.5,"passRate":0.26644736842105265,"pass_rate":0.26644736842105265,"pass_rate_percent":26.644736842105267,"passes":80,"runs":299,"runtime":349652.5,"source_flags":[],"source_harness":"codex","source_row_index":250,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":152,"tokens":null,"totalCostUsd":601.69,"totalDurationS":349652.5,"totalInputTokens":560480756,"totalOutputTokens":6366308.5,"total_cost_usd":601.69,"total_input_tokens":560480756,"total_output_tokens":6366308.5,"total_runtime_s":349652.5,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.644736842105267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-64695f2f5463c1e7bd7a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-1700f82939709812f618c036","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-low","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.16363636363636364,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":84,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":191,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[95];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-low;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":191,"metadata":{"avgScore":0.3240193712304323,"avg_score":0.3240193712304323,"avg_score_percent":32.401937123043226,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":45.35,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-low","input_tokens":29465113,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":425304,"passRate":0.16363636363636364,"pass_rate":0.16363636363636364,"pass_rate_percent":16.363636363636363,"passes":9,"runs":55,"runtime":53351,"source_flags":[],"source_harness":"codex","source_row_index":95,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":45.35,"totalDurationS":53351,"totalInputTokens":29465113,"totalOutputTokens":425304,"total_cost_usd":45.35,"total_input_tokens":29465113,"total_output_tokens":425304,"total_runtime_s":53351,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.363636363636363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d0a3cf488bb11f7c6ac3","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-15df3bf3600706c8ecd8f04c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.19736842105263158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":85,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":543,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[271];split=full/overall;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":543,"metadata":{"avgScore":0.3924651258147475,"avg_score":0.3924651258147475,"avg_score_percent":39.246512581474754,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":242.24,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":234782616,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2196372,"passRate":0.19736842105263158,"pass_rate":0.19736842105263158,"pass_rate_percent":19.736842105263158,"passes":30,"runs":139,"runtime":229751.49587275658,"source_flags":[],"source_harness":"droid","source_row_index":271,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":137,"tokens":null,"totalCostUsd":242.24,"totalDurationS":229751.49587275658,"totalInputTokens":234782616,"totalOutputTokens":2196372,"total_cost_usd":242.24,"total_input_tokens":234782616,"total_output_tokens":2196372,"total_runtime_s":229751.49587275658,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.736842105263158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ac3df608dd39c7708494","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-153f3f85f451903a2a4ef2d6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.18421052631578946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":86,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":67,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[33];split=full;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":67,"metadata":{"avgScore":0.39776664940409484,"avg_score":0.39776664940409484,"avg_score_percent":39.77666494040948,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":357.68,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":294404262,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":2853487,"passRate":0.18421052631578946,"pass_rate":0.18421052631578946,"pass_rate_percent":18.421052631578945,"passes":28,"runs":149,"runtime":270641,"source_flags":[],"source_harness":"codex","source_row_index":33,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":357.68,"totalDurationS":270641,"totalInputTokens":294404262,"totalOutputTokens":2853487,"total_cost_usd":357.68,"total_input_tokens":294404262,"total_output_tokens":2853487,"total_runtime_s":270641,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":18.421052631578945,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-89b3ec5e99e040294958","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-14d74bcfbe9b3aba2f20a81a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3283582089552239,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":87,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":389,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[194];split=full/near-term;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":389,"metadata":{"avgScore":0.6736635047623744,"avg_score":0.6736635047623744,"avg_score_percent":67.36635047623743,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":147.52,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":167219189,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1025815,"passRate":0.3283582089552239,"pass_rate":0.3283582089552239,"pass_rate_percent":32.83582089552239,"passes":22,"runs":67,"runtime":53801.491831077,"source_flags":[],"source_harness":"ale_claw","source_row_index":194,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":147.52,"totalDurationS":53801.491831077,"totalInputTokens":167219189,"totalOutputTokens":1025815,"total_cost_usd":147.52,"total_input_tokens":167219189,"total_output_tokens":1025815,"total_runtime_s":53801.491831077,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":32.83582089552239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ce2318abd721fa0eb95a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-112e725845f3c1125b4a0aae","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.17142857142857143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":88,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":785,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[392];split=linux_only;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":785,"metadata":{"avgScore":0.39445811616626686,"avg_score":0.39445811616626686,"avg_score_percent":39.445811616626685,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":194.65,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":157904128,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1928277,"passRate":0.17142857142857143,"pass_rate":0.17142857142857143,"pass_rate_percent":17.142857142857142,"passes":18,"runs":103,"runtime":95269,"source_flags":[],"source_harness":"codex","source_row_index":392,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":194.65,"totalDurationS":95269,"totalInputTokens":157904128,"totalOutputTokens":1928277,"total_cost_usd":194.65,"total_input_tokens":157904128,"total_output_tokens":1928277,"total_runtime_s":95269,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":17.142857142857142,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5460042f643ec51d74cc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-10ecf66853acc8e711f1983d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24761904761904763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":89,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":739,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[369];split=linux_only;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":739,"metadata":{"avgScore":0.480849060167781,"avg_score":0.480849060167781,"avg_score_percent":48.0849060167781,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":108.61,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":77021220,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1313359,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":103,"runtime":60234.271691576985,"source_flags":[],"source_harness":"ale_claw","source_row_index":369,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":103,"tokens":null,"totalCostUsd":108.61,"totalDurationS":60234.271691576985,"totalInputTokens":77021220,"totalOutputTokens":1313359,"total_cost_usd":108.61,"total_input_tokens":77021220,"total_output_tokens":1313359,"total_runtime_s":60234.271691576985,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.761904761904763,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9fc207972b6df391ef0c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0e75baa14d29025cc475e662","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":90,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1111,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[555];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1111,"metadata":{"avgScore":0.12697865037281922,"avg_score":0.12697865037281922,"avg_score_percent":12.697865037281922,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.16,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":101477070,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":750493,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":29,"runtime":81523,"source_flags":[],"source_harness":"codex","source_row_index":555,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":110.16,"totalDurationS":81523,"totalInputTokens":101477070,"totalOutputTokens":750493,"total_cost_usd":110.16,"total_input_tokens":101477070,"total_output_tokens":750493,"total_runtime_s":81523,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c15a37163167ed3a2d78","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0c9d68ae6d96c1325a9c384b","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"ale_claw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":91,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":151,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[75];split=full/full-spectrum;harness=ale_claw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":151,"metadata":{"avgScore":0.410514071141378,"avg_score":0.410514071141378,"avg_score_percent":41.0514071141378,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":69.9,"costSource":"db","cost_source":"db","harness":"ale_claw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":54901456,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":734763,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":54,"runtime":30454.794733498995,"source_flags":[],"source_harness":"ale_claw","source_row_index":75,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":69.9,"totalDurationS":30454.794733498995,"totalInputTokens":54901456,"totalOutputTokens":734763,"total_cost_usd":69.9,"total_input_tokens":54901456,"total_output_tokens":734763,"total_runtime_s":30454.794733498995,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.636363636363637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f47f52e13ac091cd2fa4","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-080362a9c422096de44d7c8e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0.3905472636815919,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":92,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":393,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[196];split=full/near-term;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":393,"metadata":{"avgScore":0.670526220421849,"avg_score":0.670526220421849,"avg_score_percent":67.0526220421849,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":120.93,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":162729818,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1307423.8333333333,"passRate":0.3905472636815919,"pass_rate":0.3905472636815919,"pass_rate_percent":39.054726368159194,"passes":77,"runs":198,"runtime":97188.19844049853,"source_flags":[],"source_harness":"codex","source_row_index":196,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":120.93,"totalDurationS":97188.19844049853,"totalInputTokens":162729818,"totalOutputTokens":1307423.8333333333,"total_cost_usd":120.93,"total_input_tokens":162729818,"total_output_tokens":1307423.8333333333,"total_runtime_s":97188.19844049853,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.054726368159194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-85a9831582f603a7b64d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0602c7603babdd61b872fffb","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":93,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1025,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[512];split=unlicensed/full-spectrum;harness=openclaw;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1025,"metadata":{"avgScore":0.3266237832837682,"avg_score":0.3266237832837682,"avg_score_percent":32.66237832837682,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.18,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":68495850,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":915984,"passRate":0.18,"pass_rate":0.18,"pass_rate_percent":18.0,"passes":9,"runs":47,"runtime":65810.74335963264,"source_flags":[],"source_harness":"openclaw","source_row_index":512,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":47,"tokens":null,"totalCostUsd":79.18,"totalDurationS":65810.74335963264,"totalInputTokens":68495850,"totalOutputTokens":915984,"total_cost_usd":79.18,"total_input_tokens":68495850,"total_output_tokens":915984,"total_runtime_s":65810.74335963264,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-312872b8e19cde60e23a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-045496f3e5c1aea7ae123663","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"codex","harness_variant":null,"source_harness":"codex","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":94,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1105,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[552];split=unlicensed/last-exam;harness=codex;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1105,"metadata":{"avgScore":0.1336927248629258,"avg_score":0.1336927248629258,"avg_score_percent":13.36927248629258,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":135.97,"costSource":"db+list","cost_source":"db+list","harness":"codex","harness_id":"codex","harness_variant":null,"input_tokens":169216522.5,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":987069.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":85,"runtime":87267.92330068085,"source_flags":[],"source_harness":"codex","source_row_index":552,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":30,"tokens":null,"totalCostUsd":135.97,"totalDurationS":87267.92330068085,"totalInputTokens":169216522.5,"totalOutputTokens":987069.5,"total_cost_usd":135.97,"total_input_tokens":169216522.5,"total_output_tokens":987069.5,"total_runtime_s":87267.92330068085,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c4d460dccad471009b26","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-03aa26b01bb0e5442ebaf465","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"droid","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":95,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":177,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[88];split=full/full-spectrum;harness=droid;model=gpt-5-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":177,"metadata":{"avgScore":0.3519490303044795,"avg_score":0.3519490303044795,"avg_score_percent":35.194903030447946,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":68.85,"costSource":"list","cost_source":"list","harness":"droid","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":58598314,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":792157,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":10,"runs":50,"runtime":56202.44506297966,"source_flags":[],"source_harness":"droid","source_row_index":88,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":50,"tokens":null,"totalCostUsd":68.85,"totalDurationS":56202.44506297966,"totalInputTokens":58598314,"totalOutputTokens":792157,"total_cost_usd":68.85,"total_input_tokens":58598314,"total_output_tokens":792157,"total_runtime_s":56202.44506297966,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":96,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-04b0d1bcae44127b608a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0236de841c1e7f6839fb00f8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"GPT-5.5","modelRef":"gpt-5-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"codex","harness_variant":"reasoning-medium","source_harness":"codex","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.16363636363636364,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ef1e867e055b2e2b63cb9b1cd0a5c855b4cfdf","selectionRank":96,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":189,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[94];split=full/full-spectrum;harness=codex;model=gpt-5-5;variant=reasoning-medium;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":189,"metadata":{"avgScore":0.33301050066448096,"avg_score":0.33301050066448096,"avg_score_percent":33.3010500664481,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.95,"costSource":"db","cost_source":"db","harness":"codex","harness_id":"codex","harness_variant":"reasoning-medium","input_tokens":102418661,"metric_id":"pass_rate","metric_role":"primary","model":"gpt-5-5","output_tokens":1067638,"passRate":0.16363636363636364,"pass_rate":0.16363636363636364,"pass_rate_percent":16.363636363636363,"passes":9,"runs":54,"runtime":96800,"source_flags":[],"source_harness":"codex","source_row_index":94,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":127.95,"totalDurationS":96800,"totalInputTokens":102418661,"totalOutputTokens":1067638,"total_cost_usd":127.95,"total_input_tokens":102418661,"total_output_tokens":1067638,"total_runtime_s":96800,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.363636363636363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4e8a68e7957f2a91b818","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f25b09d5c3b197a8f084199d","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5173999042242261,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1330,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[664];split=unlicensed/overall;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1330,"metadata":{"avgScore":0.5173999042242261,"avg_score":0.5173999042242261,"avg_score_percent":51.73999042242261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":468237455,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":4689597,"passRate":0.2765957446808511,"pass_rate":0.2765957446808511,"pass_rate_percent":27.659574468085108,"passes":39,"runs":140,"runtime":270855.705,"source_flags":[],"source_harness":"grok_build","source_row_index":664,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":235,"totalDurationS":270855.705,"totalInputTokens":468237455,"totalOutputTokens":4689597,"total_cost_usd":235,"total_input_tokens":468237455,"total_output_tokens":4689597,"total_runtime_s":270855.705,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":51.73999042242261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c7ddb33a1930c41ac9b9","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ec038418e9f8836e55e0c569","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.44580146295868334,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":592,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[295];split=licensed;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":592,"metadata":{"avgScore":0.44580146295868334,"avg_score":0.44580146295868334,"avg_score_percent":44.58014629586833,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":101.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":200029803,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":757770,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":2,"runs":11,"runtime":50004.460999999996,"source_flags":[],"source_harness":"grok_build","source_row_index":295,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":101.79,"totalDurationS":50004.460999999996,"totalInputTokens":200029803,"totalOutputTokens":757770,"total_cost_usd":101.79,"total_input_tokens":200029803,"total_output_tokens":757770,"total_runtime_s":50004.460999999996,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":44.58014629586833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5e562a6274232266833e","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e830857aa560628d23660dbd","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.7186616124141026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":376,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[187];split=full/near-term;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":376,"metadata":{"avgScore":0.7186616124141026,"avg_score":0.7186616124141026,"avg_score_percent":71.86616124141027,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.45,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":239613150,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":2034986,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":75263.93899999998,"source_flags":[],"source_harness":"grok_build","source_row_index":187,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":121.45,"totalDurationS":75263.93899999998,"totalInputTokens":239613150,"totalOutputTokens":2034986,"total_cost_usd":121.45,"total_input_tokens":239613150,"total_output_tokens":2034986,"total_runtime_s":75263.93899999998,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":71.86616124141027,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2b5939bee7215da19294","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e77eb528b0b87194ff5c1cd4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5122184380800093,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":14,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[6];split=full;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":14,"metadata":{"avgScore":0.5122184380800093,"avg_score":0.5122184380800093,"avg_score_percent":51.22184380800093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":336.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":668267258,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":5447367,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":320860.16600000014,"source_flags":[],"source_harness":"grok_build","source_row_index":6,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":336.79,"totalDurationS":320860.16600000014,"totalInputTokens":668267258,"totalOutputTokens":5447367,"total_cost_usd":336.79,"total_input_tokens":668267258,"total_output_tokens":5447367,"total_runtime_s":320860.16600000014,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":51.22184380800093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9172946ce6209704a4a7","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cdd9228b3c2d0d9bf9bb7cb5","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5059063643458925,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":728,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[363];split=linux_only;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":728,"metadata":{"avgScore":0.5059063643458925,"avg_score":0.5059063643458925,"avg_score_percent":50.59063643458926,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":139.15,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":247602114,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":3321922,"passRate":0.2571428571428571,"pass_rate":0.2571428571428571,"pass_rate_percent":25.71428571428571,"passes":27,"runs":104,"runtime":186556.86099999992,"source_flags":[],"source_harness":"grok_build","source_row_index":363,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":139.15,"totalDurationS":186556.86099999992,"totalInputTokens":247602114,"totalOutputTokens":3321922,"total_cost_usd":139.15,"total_input_tokens":247602114,"total_output_tokens":3321922,"total_runtime_s":186556.86099999992,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":50.59063643458926,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7622a2e87ec8c3110a62","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aae752e429b9184ec47aabbb","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4456625622278024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":132,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[65];split=full/full-spectrum;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":132,"metadata":{"avgScore":0.4456625622278024,"avg_score":0.4456625622278024,"avg_score_percent":44.56625622278024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":114.29,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":230067168,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":2071896,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":55,"runtime":107786.32800000001,"source_flags":[],"source_harness":"grok_build","source_row_index":65,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":114.29,"totalDurationS":107786.32800000001,"totalInputTokens":230067168,"totalOutputTokens":2071896,"total_cost_usd":114.29,"total_input_tokens":230067168,"total_output_tokens":2071896,"total_runtime_s":107786.32800000001,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":44.56625622278024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f5d8d2638781118a9bcf","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a5bec53d7e5381a760a949d8","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5173999042242261,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":858,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[428];split=unlicensed;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":858,"metadata":{"avgScore":0.5173999042242261,"avg_score":0.5173999042242261,"avg_score_percent":51.73999042242261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":468237455,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":4689597,"passRate":0.2765957446808511,"pass_rate":0.2765957446808511,"pass_rate_percent":27.659574468085108,"passes":39,"runs":140,"runtime":270855.705,"source_flags":[],"source_harness":"grok_build","source_row_index":428,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":235,"totalDurationS":270855.705,"totalInputTokens":468237455,"totalOutputTokens":4689597,"total_cost_usd":235,"total_input_tokens":468237455,"total_output_tokens":4689597,"total_runtime_s":270855.705,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":51.73999042242261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bacb414efdc11402903f","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-657b5983c40cac1df2402ec4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1967489930133485,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1084,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[541];split=unlicensed/last-exam;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1084,"metadata":{"avgScore":0.1967489930133485,"avg_score":0.1967489930133485,"avg_score_percent":19.67489930133485,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":59.19,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":118841427,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":1292528,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":29,"runtime":132480.78900000002,"source_flags":[],"source_harness":"grok_build","source_row_index":541,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":59.19,"totalDurationS":132480.78900000002,"totalInputTokens":118841427,"totalOutputTokens":1292528,"total_cost_usd":59.19,"total_input_tokens":118841427,"total_output_tokens":1292528,"total_runtime_s":132480.78900000002,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":19.67489930133485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-896186280ccd0048e09d","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-55747b6e64f8fc164d8fe22b","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23700752323542038,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":240,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[119];split=full/last-exam;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":240,"metadata":{"avgScore":0.23700752323542038,"avg_score":0.23700752323542038,"avg_score_percent":23.700752323542037,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":114.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":223738865,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":1749767,"passRate":0.10526315789473684,"pass_rate":0.10526315789473684,"pass_rate_percent":10.526315789473683,"passes":4,"runs":37,"runtime":171781.49400000004,"source_flags":[],"source_harness":"grok_build","source_row_index":119,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":114.79,"totalDurationS":171781.49400000004,"totalInputTokens":223738865,"totalOutputTokens":1749767,"total_cost_usd":114.79,"total_input_tokens":223738865,"total_output_tokens":1749767,"total_runtime_s":171781.49400000004,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.700752323542037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d2c4641ee40d1e8a4859","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4ae1c5c409b3b18b1b5d6dce","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.4247048184505826,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":978,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[488];split=unlicensed/full-spectrum;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":978,"metadata":{"avgScore":0.4247048184505826,"avg_score":0.4247048184505826,"avg_score_percent":42.47048184505826,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":67.58,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":133916121,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":1752563,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":50,"runtime":78269.59600000002,"source_flags":[],"source_harness":"grok_build","source_row_index":488,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":67.58,"totalDurationS":78269.59600000002,"totalInputTokens":133916121,"totalOutputTokens":1752563,"total_cost_usd":67.58,"total_input_tokens":133916121,"total_output_tokens":1752563,"total_runtime_s":78269.59600000002,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":42.47048184505826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-97ef2cf1e90fe3ee26df","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2174daac414e6d48624f4f6e","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.5122184380800093,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":486,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[242];split=full/overall;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":486,"metadata":{"avgScore":0.5122184380800093,"avg_score":0.5122184380800093,"avg_score_percent":51.22184380800093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":336.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":668267258,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":5447367,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":320860.16600000014,"source_flags":[],"source_harness":"grok_build","source_row_index":242,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":336.79,"totalDurationS":320860.16600000014,"totalInputTokens":668267258,"totalOutputTokens":5447367,"total_cost_usd":336.79,"total_input_tokens":668267258,"total_output_tokens":5447367,"total_runtime_s":320860.16600000014,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":51.22184380800093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e5dce9246762fb19233","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0883694ab4b64b561aa04dc7","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.7186616124141026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e6469a42d092ddd2c951793339c38841c4e0b5","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1212,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[605];split=unlicensed/near-term;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1212,"metadata":{"avgScore":0.7186616124141026,"avg_score":0.7186616124141026,"avg_score_percent":71.86616124141027,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.45,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":239613150,"metric_id":"avg_score","metric_role":"secondary","model":"grok-4-5","output_tokens":2034986,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":75263.93899999998,"source_flags":[],"source_harness":"grok_build","source_row_index":605,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":121.45,"totalDurationS":75263.93899999998,"totalInputTokens":239613150,"totalOutputTokens":2034986,"total_cost_usd":121.45,"total_input_tokens":239613150,"total_output_tokens":2034986,"total_runtime_s":75263.93899999998,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":71.86616124141027,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7298a595d497aea432b7","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f0d77113d2de9cf754e489fb","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":13,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[6];split=full;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":13,"metadata":{"avgScore":0.5122184380800093,"avg_score":0.5122184380800093,"avg_score_percent":51.22184380800093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":336.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":668267258,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":5447367,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":320860.16600000014,"source_flags":[],"source_harness":"grok_build","source_row_index":6,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":336.79,"totalDurationS":320860.16600000014,"totalInputTokens":668267258,"totalOutputTokens":5447367,"total_cost_usd":336.79,"total_input_tokens":668267258,"total_output_tokens":5447367,"total_runtime_s":320860.16600000014,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1c2db2616a7310d3412a","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dcf3ed0dc8f6c973e2373ca4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":977,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[488];split=unlicensed/full-spectrum;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":977,"metadata":{"avgScore":0.4247048184505826,"avg_score":0.4247048184505826,"avg_score_percent":42.47048184505826,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":67.58,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":133916121,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":1752563,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":50,"runtime":78269.59600000002,"source_flags":[],"source_harness":"grok_build","source_row_index":488,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":67.58,"totalDurationS":78269.59600000002,"totalInputTokens":133916121,"totalOutputTokens":1752563,"total_cost_usd":67.58,"total_input_tokens":133916121,"total_output_tokens":1752563,"total_runtime_s":78269.59600000002,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8b23434a75ecab0a4102","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b5d540cf7f40be4e09f52061","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":375,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[187];split=full/near-term;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":375,"metadata":{"avgScore":0.7186616124141026,"avg_score":0.7186616124141026,"avg_score_percent":71.86616124141027,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.45,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":239613150,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":2034986,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":75263.93899999998,"source_flags":[],"source_harness":"grok_build","source_row_index":187,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":121.45,"totalDurationS":75263.93899999998,"totalInputTokens":239613150,"totalOutputTokens":2034986,"total_cost_usd":121.45,"total_input_tokens":239613150,"total_output_tokens":2034986,"total_runtime_s":75263.93899999998,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c1a22d1efe2a491d3d55","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b464be43da3edf29301b68c9","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1211,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[605];split=unlicensed/near-term;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1211,"metadata":{"avgScore":0.7186616124141026,"avg_score":0.7186616124141026,"avg_score_percent":71.86616124141027,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.45,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":239613150,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":2034986,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":75263.93899999998,"source_flags":[],"source_harness":"grok_build","source_row_index":605,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":121.45,"totalDurationS":75263.93899999998,"totalInputTokens":239613150,"totalOutputTokens":2034986,"total_cost_usd":121.45,"total_input_tokens":239613150,"total_output_tokens":2034986,"total_runtime_s":75263.93899999998,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-18a02fd52e0351fb1738","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aa61b483c2fcc78de8561aae","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1083,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[541];split=unlicensed/last-exam;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1083,"metadata":{"avgScore":0.1967489930133485,"avg_score":0.1967489930133485,"avg_score_percent":19.67489930133485,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":59.19,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":118841427,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":1292528,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":29,"runtime":132480.78900000002,"source_flags":[],"source_harness":"grok_build","source_row_index":541,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":59.19,"totalDurationS":132480.78900000002,"totalInputTokens":118841427,"totalOutputTokens":1292528,"total_cost_usd":59.19,"total_input_tokens":118841427,"total_output_tokens":1292528,"total_runtime_s":132480.78900000002,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e4b66d5764a30cf4c2e0","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-65bd62a9baf4522d41479c36","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2765957446808511,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1329,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[664];split=unlicensed/overall;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1329,"metadata":{"avgScore":0.5173999042242261,"avg_score":0.5173999042242261,"avg_score_percent":51.73999042242261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":468237455,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":4689597,"passRate":0.2765957446808511,"pass_rate":0.2765957446808511,"pass_rate_percent":27.659574468085108,"passes":39,"runs":140,"runtime":270855.705,"source_flags":[],"source_harness":"grok_build","source_row_index":664,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":235,"totalDurationS":270855.705,"totalInputTokens":468237455,"totalOutputTokens":4689597,"total_cost_usd":235,"total_input_tokens":468237455,"total_output_tokens":4689597,"total_runtime_s":270855.705,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":27.659574468085108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e74d918e6893c3956ecb","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-53820e84074364c6ba721235","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":131,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[65];split=full/full-spectrum;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":131,"metadata":{"avgScore":0.4456625622278024,"avg_score":0.4456625622278024,"avg_score_percent":44.56625622278024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":114.29,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":230067168,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":2071896,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":55,"runtime":107786.32800000001,"source_flags":[],"source_harness":"grok_build","source_row_index":65,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":114.29,"totalDurationS":107786.32800000001,"totalInputTokens":230067168,"totalOutputTokens":2071896,"total_cost_usd":114.29,"total_input_tokens":230067168,"total_output_tokens":2071896,"total_runtime_s":107786.32800000001,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.636363636363637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b1671661797dcd8f420a","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4c6834b678a837f5caa472a9","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10526315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":239,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[119];split=full/last-exam;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":239,"metadata":{"avgScore":0.23700752323542038,"avg_score":0.23700752323542038,"avg_score_percent":23.700752323542037,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":114.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":223738865,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":1749767,"passRate":0.10526315789473684,"pass_rate":0.10526315789473684,"pass_rate_percent":10.526315789473683,"passes":4,"runs":37,"runtime":171781.49400000004,"source_flags":[],"source_harness":"grok_build","source_row_index":119,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":114.79,"totalDurationS":171781.49400000004,"totalInputTokens":223738865,"totalOutputTokens":1749767,"total_cost_usd":114.79,"total_input_tokens":223738865,"total_output_tokens":1749767,"total_runtime_s":171781.49400000004,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.526315789473683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-db0de014809b842d45fb","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2efc6a326d80e268f6025a30","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2571428571428571,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":727,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[363];split=linux_only;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":727,"metadata":{"avgScore":0.5059063643458925,"avg_score":0.5059063643458925,"avg_score_percent":50.59063643458926,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":139.15,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":247602114,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":3321922,"passRate":0.2571428571428571,"pass_rate":0.2571428571428571,"pass_rate_percent":25.71428571428571,"passes":27,"runs":104,"runtime":186556.86099999992,"source_flags":[],"source_harness":"grok_build","source_row_index":363,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":139.15,"totalDurationS":186556.86099999992,"totalInputTokens":247602114,"totalOutputTokens":3321922,"total_cost_usd":139.15,"total_input_tokens":247602114,"total_output_tokens":3321922,"total_runtime_s":186556.86099999992,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":25.71428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5374019260b5efee3177","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2d5b2bd9b82de8e2066a441c","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2765957446808511,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":857,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[428];split=unlicensed;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":857,"metadata":{"avgScore":0.5173999042242261,"avg_score":0.5173999042242261,"avg_score_percent":51.73999042242261,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":235,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":468237455,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":4689597,"passRate":0.2765957446808511,"pass_rate":0.2765957446808511,"pass_rate_percent":27.659574468085108,"passes":39,"runs":140,"runtime":270855.705,"source_flags":[],"source_harness":"grok_build","source_row_index":428,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":140,"tokens":null,"totalCostUsd":235,"totalDurationS":270855.705,"totalInputTokens":468237455,"totalOutputTokens":4689597,"total_cost_usd":235,"total_input_tokens":468237455,"total_output_tokens":4689597,"total_runtime_s":270855.705,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":27.659574468085108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a997823fa57d6678589f","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-15f178a26b31f7742b8af121","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":485,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[242];split=full/overall;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":485,"metadata":{"avgScore":0.5122184380800093,"avg_score":0.5122184380800093,"avg_score_percent":51.22184380800093,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":336.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":668267258,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":5447367,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":151,"runtime":320860.16600000014,"source_flags":[],"source_harness":"grok_build","source_row_index":242,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":151,"tokens":null,"totalCostUsd":336.79,"totalDurationS":320860.16600000014,"totalInputTokens":668267258,"totalOutputTokens":5447367,"total_cost_usd":336.79,"total_input_tokens":668267258,"total_output_tokens":5447367,"total_runtime_s":320860.16600000014,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e73c83e4df490ed38453","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0cce4a1491b2d433ae17f837","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of Grok 4.5; harness remains in the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Grok 4.5","modelRef":"grok-4-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":"reasoning-high","source_harness":"grok_build","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8efefaec500e85be184c78cbd429fb8604094fae","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":591,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[295];split=licensed;harness=grok_build;model=grok-4-5;variant=reasoning-high;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":591,"metadata":{"avgScore":0.44580146295868334,"avg_score":0.44580146295868334,"avg_score_percent":44.58014629586833,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":101.79,"costSource":"db","cost_source":"db","harness":"grok_build","harness_id":"agents-last-exam","harness_variant":"reasoning-high","input_tokens":200029803,"metric_id":"pass_rate","metric_role":"primary","model":"grok-4-5","output_tokens":757770,"passRate":0.18181818181818182,"pass_rate":0.18181818181818182,"pass_rate_percent":18.181818181818183,"passes":2,"runs":11,"runtime":50004.460999999996,"source_flags":[],"source_harness":"grok_build","source_row_index":295,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":101.79,"totalDurationS":50004.460999999996,"totalInputTokens":200029803,"totalOutputTokens":757770,"total_cost_usd":101.79,"total_input_tokens":200029803,"total_output_tokens":757770,"total_runtime_s":50004.460999999996,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.181818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0bf91d257646c4d3cce8","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e8e8319929c08f91acdefaa3","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.18828448362311412,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1062,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[530];split=unlicensed/full-spectrum;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1062,"metadata":{"avgScore":0.18828448362311412,"avg_score":0.18828448362311412,"avg_score_percent":18.828448362311413,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":39.81,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":92029776,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":2361976.5,"passRate":0.07,"pass_rate":0.07,"pass_rate_percent":7.000000000000001,"passes":6,"runs":82,"runtime":204363.24344928894,"source_flags":[],"source_harness":"openclaw","source_row_index":530,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":39.81,"totalDurationS":204363.24344928894,"totalInputTokens":92029776,"totalOutputTokens":2361976.5,"total_cost_usd":39.81,"total_input_tokens":92029776,"total_output_tokens":2361976.5,"total_runtime_s":204363.24344928894,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.828448362311413,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fda2d3f165f442549ec2","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e51fd89df2b8580c0ce1f70c","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.01640722406550101,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":348,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[173];split=full/last-exam;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":348,"metadata":{"avgScore":0.01640722406550101,"avg_score":0.01640722406550101,"avg_score_percent":1.640722406550101,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.87,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":88077257,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":1829053.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":48,"runtime":218740.34341101287,"source_flags":[],"source_harness":"openclaw","source_row_index":173,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":38.87,"totalDurationS":218740.34341101287,"totalInputTokens":88077257,"totalOutputTokens":1829053.5,"total_cost_usd":38.87,"total_input_tokens":88077257,"total_output_tokens":1829053.5,"total_runtime_s":218740.34341101287,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":1.640722406550101,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1b7141d1f84a5ac4624","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cee40f467499f39db1d3076d","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1820768032937401,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":226,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[112];split=full/full-spectrum;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":226,"metadata":{"avgScore":0.1820768032937401,"avg_score":0.1820768032937401,"avg_score_percent":18.20768032937401,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":41.17,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":93552781,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":2428431.5,"passRate":0.06363636363636363,"pass_rate":0.06363636363636363,"pass_rate_percent":6.363636363636363,"passes":6,"runs":87,"runtime":264884.5961624289,"source_flags":[],"source_harness":"openclaw","source_row_index":112,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":41.17,"totalDurationS":264884.5961624289,"totalInputTokens":93552781,"totalOutputTokens":2428431.5,"total_cost_usd":41.17,"total_input_tokens":93552781,"total_output_tokens":2428431.5,"total_runtime_s":264884.5961624289,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":18.20768032937401,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9210b2d024526ec6a5b0","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cd8049275c928980752933e0","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2123825941816749,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":826,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[412];split=linux_only;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":826,"metadata":{"avgScore":0.2123825941816749,"avg_score":0.2123825941816749,"avg_score_percent":21.23825941816749,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":90.85,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":222945194.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":4427429,"passRate":0.08095238095238096,"pass_rate":0.08095238095238096,"pass_rate_percent":8.095238095238095,"passes":16,"runs":188,"runtime":322265.0739102186,"source_flags":[],"source_harness":"openclaw","source_row_index":412,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":90.85,"totalDurationS":322265.0739102186,"totalInputTokens":222945194.5,"totalOutputTokens":4427429,"total_cost_usd":90.85,"total_input_tokens":222945194.5,"total_output_tokens":4427429,"total_runtime_s":322265.0739102186,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.23825941816749,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e27ac0f1722b1fb85fdd","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cbe911c59ff0103cd13e0965","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2292968817611259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1416,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[707];split=unlicensed/overall;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1416,"metadata":{"avgScore":0.2292968817611259,"avg_score":0.2292968817611259,"avg_score_percent":22.929688176112588,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.61,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":293251082.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":5966124,"passRate":0.09929078014184398,"pass_rate":0.09929078014184398,"pass_rate_percent":9.929078014184398,"passes":25,"runs":234,"runtime":563409.5122970284,"source_flags":[],"source_harness":"openclaw","source_row_index":707,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":121.61,"totalDurationS":563409.5122970284,"totalInputTokens":293251082.5,"totalOutputTokens":5966124,"total_cost_usd":121.61,"total_input_tokens":293251082.5,"total_output_tokens":5966124,"total_runtime_s":563409.5122970284,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.929688176112588,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f57422aecb4e990ebf35","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b3f6964a5c62c1ac2f575c47","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3512706887636276,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":460,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[229];split=full/near-term;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":460,"metadata":{"avgScore":0.3512706887636276,"avg_score":0.3512706887636276,"avg_score_percent":35.12706887636276,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.55,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":118462415.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":2057036.5,"passRate":0.15671641791044777,"pass_rate":0.15671641791044777,"pass_rate_percent":15.671641791044777,"passes":19,"runs":119,"runtime":219933.07505434786,"source_flags":[],"source_harness":"openclaw","source_row_index":229,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":46.55,"totalDurationS":219933.07505434786,"totalInputTokens":118462415.5,"totalOutputTokens":2057036.5,"total_cost_usd":46.55,"total_input_tokens":118462415.5,"total_output_tokens":2057036.5,"total_runtime_s":219933.07505434786,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.12706887636276,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1cbe4b862ff144b7b43","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-97855e2f3517c2b9bc8c7d0f","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21665039689683385,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":580,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[289];split=full/overall;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":580,"metadata":{"avgScore":0.21665039689683385,"avg_score":0.21665039689683385,"avg_score_percent":21.665039689683386,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":123.45,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":295931464.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":6043559,"passRate":0.09210526315789473,"pass_rate":0.09210526315789473,"pass_rate_percent":9.210526315789473,"passes":25,"runs":241,"runtime":659868.3114792772,"source_flags":[],"source_harness":"openclaw","source_row_index":289,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":123.45,"totalDurationS":659868.3114792772,"totalInputTokens":295931464.5,"totalOutputTokens":6043559,"total_cost_usd":123.45,"total_input_tokens":295931464.5,"total_output_tokens":6043559,"total_runtime_s":659868.3114792772,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.665039689683386,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d05e110b316bf0cb8f08","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-72fca6480d7f90a313399d6c","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.020782483816301277,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1184,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[591];split=unlicensed/last-exam;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1184,"metadata":{"avgScore":0.020782483816301277,"avg_score":0.020782483816301277,"avg_score_percent":2.078248381630128,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.24,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86713583,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":1803204.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":44,"runtime":162660.51901915795,"source_flags":[],"source_harness":"openclaw","source_row_index":591,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":38.24,"totalDurationS":162660.51901915795,"totalInputTokens":86713583,"totalOutputTokens":1803204.5,"total_cost_usd":38.24,"total_input_tokens":86713583,"total_output_tokens":1803204.5,"total_runtime_s":162660.51901915795,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":2.078248381630128,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3f58c0a46b966402f48d","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6b286d939e104f758b5e0a8d","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3512706887636276,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1296,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[647];split=unlicensed/near-term;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1296,"metadata":{"avgScore":0.3512706887636276,"avg_score":0.3512706887636276,"avg_score_percent":35.12706887636276,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.55,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":118462415.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":2057036.5,"passRate":0.15671641791044777,"pass_rate":0.15671641791044777,"pass_rate_percent":15.671641791044777,"passes":19,"runs":119,"runtime":219933.07505434786,"source_flags":[],"source_harness":"openclaw","source_row_index":647,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":46.55,"totalDurationS":219933.07505434786,"totalInputTokens":118462415.5,"totalOutputTokens":2057036.5,"total_cost_usd":46.55,"total_input_tokens":118462415.5,"total_output_tokens":2057036.5,"total_runtime_s":219933.07505434786,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.12706887636276,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e38ffa88caf59a2fad6a","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4e43b3bd1c978e8d60babfd1","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.05454545454545454,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":690,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[344];split=licensed;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":690,"metadata":{"avgScore":0.05454545454545454,"avg_score":0.05454545454545454,"avg_score_percent":5.454545454545454,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1.85,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2680382,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":77435,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":96458.79918224877,"source_flags":[],"source_harness":"openclaw","source_row_index":344,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":1.85,"totalDurationS":96458.79918224877,"totalInputTokens":2680382,"totalOutputTokens":77435,"total_cost_usd":1.85,"total_input_tokens":2680382,"total_output_tokens":77435,"total_runtime_s":96458.79918224877,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":5.454545454545454,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a0e6b98d93595ab96a19","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4c36257f7de1b29a8bfec790","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2292968817611259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":944,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[471];split=unlicensed;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":944,"metadata":{"avgScore":0.2292968817611259,"avg_score":0.2292968817611259,"avg_score_percent":22.929688176112588,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.61,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":293251082.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":5966124,"passRate":0.09929078014184398,"pass_rate":0.09929078014184398,"pass_rate_percent":9.929078014184398,"passes":25,"runs":234,"runtime":563409.5122970284,"source_flags":[],"source_harness":"openclaw","source_row_index":471,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":121.61,"totalDurationS":563409.5122970284,"totalInputTokens":293251082.5,"totalOutputTokens":5966124,"total_cost_usd":121.61,"total_input_tokens":293251082.5,"total_output_tokens":5966124,"total_runtime_s":563409.5122970284,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.929688176112588,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1b8123be24d434a0d06","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1ab8076d8c8de79c3be3208f","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21665039689683385,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10ad86acf1c66a889b562c0a6faa7b547537fa2c","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":108,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[53];split=full;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":108,"metadata":{"avgScore":0.21665039689683385,"avg_score":0.21665039689683385,"avg_score_percent":21.665039689683386,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":123.45,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":295931464.5,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k2-6","output_tokens":6043559,"passRate":0.09210526315789473,"pass_rate":0.09210526315789473,"pass_rate_percent":9.210526315789473,"passes":25,"runs":241,"runtime":659868.3114792772,"source_flags":[],"source_harness":"openclaw","source_row_index":53,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":123.45,"totalDurationS":659868.3114792772,"totalInputTokens":295931464.5,"totalOutputTokens":6043559,"total_cost_usd":123.45,"total_input_tokens":295931464.5,"total_output_tokens":6043559,"total_runtime_s":659868.3114792772,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.665039689683386,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f4f8ee58515ad2a31c56","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c99abe0cdd467065c533bfaf","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.15671641791044777,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1295,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[647];split=unlicensed/near-term;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1295,"metadata":{"avgScore":0.3512706887636276,"avg_score":0.3512706887636276,"avg_score_percent":35.12706887636276,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.55,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":118462415.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":2057036.5,"passRate":0.15671641791044777,"pass_rate":0.15671641791044777,"pass_rate_percent":15.671641791044777,"passes":19,"runs":119,"runtime":219933.07505434786,"source_flags":[],"source_harness":"openclaw","source_row_index":647,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":46.55,"totalDurationS":219933.07505434786,"totalInputTokens":118462415.5,"totalOutputTokens":2057036.5,"total_cost_usd":46.55,"total_input_tokens":118462415.5,"total_output_tokens":2057036.5,"total_runtime_s":219933.07505434786,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.671641791044777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-42c855055349d1e98446","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c576f6d162457339248d6911","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09210526315789473,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":579,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[289];split=full/overall;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":579,"metadata":{"avgScore":0.21665039689683385,"avg_score":0.21665039689683385,"avg_score_percent":21.665039689683386,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":123.45,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":295931464.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":6043559,"passRate":0.09210526315789473,"pass_rate":0.09210526315789473,"pass_rate_percent":9.210526315789473,"passes":25,"runs":241,"runtime":659868.3114792772,"source_flags":[],"source_harness":"openclaw","source_row_index":289,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":123.45,"totalDurationS":659868.3114792772,"totalInputTokens":295931464.5,"totalOutputTokens":6043559,"total_cost_usd":123.45,"total_input_tokens":295931464.5,"total_output_tokens":6043559,"total_runtime_s":659868.3114792772,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.210526315789473,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a0f9cc743f341a9568ac","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c15381d3a086a5289cb3cc95","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":347,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[173];split=full/last-exam;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":347,"metadata":{"avgScore":0.01640722406550101,"avg_score":0.01640722406550101,"avg_score_percent":1.640722406550101,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.87,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":88077257,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":1829053.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":48,"runtime":218740.34341101287,"source_flags":[],"source_harness":"openclaw","source_row_index":173,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":38.87,"totalDurationS":218740.34341101287,"totalInputTokens":88077257,"totalOutputTokens":1829053.5,"total_cost_usd":38.87,"total_input_tokens":88077257,"total_output_tokens":1829053.5,"total_runtime_s":218740.34341101287,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-062168bc276938585b89","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9d4db1cb4dbad058b1c3c099","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09929078014184398,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":943,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[471];split=unlicensed;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":943,"metadata":{"avgScore":0.2292968817611259,"avg_score":0.2292968817611259,"avg_score_percent":22.929688176112588,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.61,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":293251082.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":5966124,"passRate":0.09929078014184398,"pass_rate":0.09929078014184398,"pass_rate_percent":9.929078014184398,"passes":25,"runs":234,"runtime":563409.5122970284,"source_flags":[],"source_harness":"openclaw","source_row_index":471,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":121.61,"totalDurationS":563409.5122970284,"totalInputTokens":293251082.5,"totalOutputTokens":5966124,"total_cost_usd":121.61,"total_input_tokens":293251082.5,"total_output_tokens":5966124,"total_runtime_s":563409.5122970284,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.929078014184398,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-926e419d676338ecaf8f","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8e227924ff33f5881718031b","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1183,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[591];split=unlicensed/last-exam;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1183,"metadata":{"avgScore":0.020782483816301277,"avg_score":0.020782483816301277,"avg_score_percent":2.078248381630128,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.24,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86713583,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":1803204.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":44,"runtime":162660.51901915795,"source_flags":[],"source_harness":"openclaw","source_row_index":591,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":38.24,"totalDurationS":162660.51901915795,"totalInputTokens":86713583,"totalOutputTokens":1803204.5,"total_cost_usd":38.24,"total_input_tokens":86713583,"total_output_tokens":1803204.5,"total_runtime_s":162660.51901915795,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c1f0464137fa440ea117","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6b674bba8e058655a67d65a3","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.07,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1061,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[530];split=unlicensed/full-spectrum;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1061,"metadata":{"avgScore":0.18828448362311412,"avg_score":0.18828448362311412,"avg_score_percent":18.828448362311413,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":39.81,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":92029776,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":2361976.5,"passRate":0.07,"pass_rate":0.07,"pass_rate_percent":7.000000000000001,"passes":6,"runs":82,"runtime":204363.24344928894,"source_flags":[],"source_harness":"openclaw","source_row_index":530,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":39.81,"totalDurationS":204363.24344928894,"totalInputTokens":92029776,"totalOutputTokens":2361976.5,"total_cost_usd":39.81,"total_input_tokens":92029776,"total_output_tokens":2361976.5,"total_runtime_s":204363.24344928894,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-63e5a36aa321c1f9d3a6","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-418c399d612cb96c5205890d","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08095238095238096,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":825,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[412];split=linux_only;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":825,"metadata":{"avgScore":0.2123825941816749,"avg_score":0.2123825941816749,"avg_score_percent":21.23825941816749,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":90.85,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":222945194.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":4427429,"passRate":0.08095238095238096,"pass_rate":0.08095238095238096,"pass_rate_percent":8.095238095238095,"passes":16,"runs":188,"runtime":322265.0739102186,"source_flags":[],"source_harness":"openclaw","source_row_index":412,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":90.85,"totalDurationS":322265.0739102186,"totalInputTokens":222945194.5,"totalOutputTokens":4427429,"total_cost_usd":90.85,"total_input_tokens":222945194.5,"total_output_tokens":4427429,"total_runtime_s":322265.0739102186,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.095238095238095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-00aff5db30f4a31bc6da","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3adda6de68785bcc8aa9b8f2","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06363636363636363,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":225,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[112];split=full/full-spectrum;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":225,"metadata":{"avgScore":0.1820768032937401,"avg_score":0.1820768032937401,"avg_score_percent":18.20768032937401,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":41.17,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":93552781,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":2428431.5,"passRate":0.06363636363636363,"pass_rate":0.06363636363636363,"pass_rate_percent":6.363636363636363,"passes":6,"runs":87,"runtime":264884.5961624289,"source_flags":[],"source_harness":"openclaw","source_row_index":112,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":41.17,"totalDurationS":264884.5961624289,"totalInputTokens":93552781,"totalOutputTokens":2428431.5,"total_cost_usd":41.17,"total_input_tokens":93552781,"total_output_tokens":2428431.5,"total_runtime_s":264884.5961624289,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.363636363636363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c367bbc13ed2b4fb0135","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1f566c6090a3d9dc3a567b32","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.15671641791044777,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":459,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[229];split=full/near-term;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":459,"metadata":{"avgScore":0.3512706887636276,"avg_score":0.3512706887636276,"avg_score_percent":35.12706887636276,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":46.55,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":118462415.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":2057036.5,"passRate":0.15671641791044777,"pass_rate":0.15671641791044777,"pass_rate_percent":15.671641791044777,"passes":19,"runs":119,"runtime":219933.07505434786,"source_flags":[],"source_harness":"openclaw","source_row_index":229,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":65,"tokens":null,"totalCostUsd":46.55,"totalDurationS":219933.07505434786,"totalInputTokens":118462415.5,"totalOutputTokens":2057036.5,"total_cost_usd":46.55,"total_input_tokens":118462415.5,"total_output_tokens":2057036.5,"total_runtime_s":219933.07505434786,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":15.671641791044777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-99a283ccbce855ed4828","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1dd0045adbcd3bc038780c4a","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":689,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[344];split=licensed;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":689,"metadata":{"avgScore":0.05454545454545454,"avg_score":0.05454545454545454,"avg_score_percent":5.454545454545454,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1.85,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2680382,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":77435,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":96458.79918224877,"source_flags":[],"source_harness":"openclaw","source_row_index":344,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":1.85,"totalDurationS":96458.79918224877,"totalInputTokens":2680382,"totalOutputTokens":77435,"total_cost_usd":1.85,"total_input_tokens":2680382,"total_output_tokens":77435,"total_runtime_s":96458.79918224877,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3a91b8e6e75491470d07","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-14852aa8f327773f92ce78b6","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09210526315789473,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":107,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[53];split=full;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":107,"metadata":{"avgScore":0.21665039689683385,"avg_score":0.21665039689683385,"avg_score_percent":21.665039689683386,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":123.45,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":295931464.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":6043559,"passRate":0.09210526315789473,"pass_rate":0.09210526315789473,"pass_rate_percent":9.210526315789473,"passes":25,"runs":241,"runtime":659868.3114792772,"source_flags":[],"source_harness":"openclaw","source_row_index":53,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":144,"tokens":null,"totalCostUsd":123.45,"totalDurationS":659868.3114792772,"totalInputTokens":295931464.5,"totalOutputTokens":6043559,"total_cost_usd":123.45,"total_input_tokens":295931464.5,"total_output_tokens":6043559,"total_runtime_s":659868.3114792772,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.210526315789473,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d81e8fcde58e9a6f08ec","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-07aa390d356533c9fd4996d9","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K2.6","modelRef":"kimi-k2-6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09929078014184398,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"835287fe53eb2fe8736f3c5f4b9f5885d0cce165","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1415,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[707];split=unlicensed/overall;harness=openclaw;model=kimi-k2-6;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1415,"metadata":{"avgScore":0.2292968817611259,"avg_score":0.2292968817611259,"avg_score_percent":22.929688176112588,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":121.61,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":293251082.5,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k2-6","output_tokens":5966124,"passRate":0.09929078014184398,"pass_rate":0.09929078014184398,"pass_rate_percent":9.929078014184398,"passes":25,"runs":234,"runtime":563409.5122970284,"source_flags":[],"source_harness":"openclaw","source_row_index":707,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":137,"tokens":null,"totalCostUsd":121.61,"totalDurationS":563409.5122970284,"totalInputTokens":293251082.5,"totalOutputTokens":5966124,"total_cost_usd":121.61,"total_input_tokens":293251082.5,"total_output_tokens":5966124,"total_runtime_s":563409.5122970284,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.929078014184398,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8c2a2bdb2ed995937cc0","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-fb0a9218c492c80e4a74e664","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.515832318778052,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":12,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[5];split=full;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":12,"metadata":{"avgScore":0.515832318778052,"avg_score":0.515832318778052,"avg_score_percent":51.583231877805204,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":605.78,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":1402173402,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":7479641,"passRate":0.28289473684210525,"pass_rate":0.28289473684210525,"pass_rate_percent":28.289473684210524,"passes":43,"runs":149,"runtime":669880.985,"source_flags":[],"source_harness":"kimi_code","source_row_index":5,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":605.78,"totalDurationS":669880.985,"totalInputTokens":1402173402,"totalOutputTokens":7479641,"total_cost_usd":605.78,"total_input_tokens":1402173402,"total_output_tokens":7479641,"total_runtime_s":669880.985,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.583231877805204,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-22a09159d9be1fdec760","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e4187cb193fa4b698bfdd707","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7159801555439591,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1214,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[606];split=unlicensed/near-term;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1214,"metadata":{"avgScore":0.7159801555439591,"avg_score":0.7159801555439591,"avg_score_percent":71.59801555439591,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":145.05,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":270158643,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2722015,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":239761.4640000001,"source_flags":[],"source_harness":"claude_code","source_row_index":606,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":145.05,"totalDurationS":239761.4640000001,"totalInputTokens":270158643,"totalOutputTokens":2722015,"total_cost_usd":145.05,"total_input_tokens":270158643,"total_output_tokens":2722015,"total_runtime_s":239761.4640000001,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":71.59801555439591,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2f7797059d832470b1b3","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e2903a242fe069a68fed0d59","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2049303255658983,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":246,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[122];split=full/last-exam;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":246,"metadata":{"avgScore":0.2049303255658983,"avg_score":0.2049303255658983,"avg_score_percent":20.49303255658983,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":200.05,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":437612423,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2624214,"passRate":0.07894736842105263,"pass_rate":0.07894736842105263,"pass_rate_percent":7.894736842105263,"passes":3,"runs":36,"runtime":363134.68600000005,"source_flags":[],"source_harness":"claude_code","source_row_index":122,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":200.05,"totalDurationS":363134.68600000005,"totalInputTokens":437612423,"totalOutputTokens":2624214,"total_cost_usd":200.05,"total_input_tokens":437612423,"total_output_tokens":2624214,"total_runtime_s":363134.68600000005,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.49303255658983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4136aea52943c4b7c718","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-cd54bd1303748c364b4ac0f7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5278637305133695,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":714,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[356];split=linux_only;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":714,"metadata":{"avgScore":0.5278637305133695,"avg_score":0.5278637305133695,"avg_score_percent":52.78637305133695,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":251.07,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":465187955,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":4686442,"passRate":0.2761904761904762,"pass_rate":0.2761904761904762,"pass_rate_percent":27.61904761904762,"passes":29,"runs":102,"runtime":393848.4819999999,"source_flags":[],"source_harness":"claude_code","source_row_index":356,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":251.07,"totalDurationS":393848.4819999999,"totalInputTokens":465187955,"totalOutputTokens":4686442,"total_cost_usd":251.07,"total_input_tokens":465187955,"total_output_tokens":4686442,"total_runtime_s":393848.4819999999,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.78637305133695,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-79b5debc08c4e4cde816","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-c442c47a1eb3e0889ded42e6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6968586979044264,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1220,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[609];split=unlicensed/near-term;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1220,"metadata":{"avgScore":0.6968586979044264,"avg_score":0.6968586979044264,"avg_score_percent":69.68586979044264,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":129.35,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":256675448,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2162427,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":66,"runtime":182895.38799999998,"source_flags":[],"source_harness":"kimi_code","source_row_index":609,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":129.35,"totalDurationS":182895.38799999998,"totalInputTokens":256675448,"totalOutputTokens":2162427,"total_cost_usd":129.35,"total_input_tokens":256675448,"total_output_tokens":2162427,"total_runtime_s":182895.38799999998,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":69.68586979044264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3d2e5004c5043c6a7f1d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-bd5b0a4977c5e9963d1dc380","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5072487812208666,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":16,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[7];split=full;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":16,"metadata":{"avgScore":0.5072487812208666,"avg_score":0.5072487812208666,"avg_score_percent":50.72487812208666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":463.87,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":950614474,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":7394400,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":149,"runtime":775905.225,"source_flags":[],"source_harness":"claude_code","source_row_index":7,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":463.87,"totalDurationS":775905.225,"totalInputTokens":950614474,"totalOutputTokens":7394400,"total_cost_usd":463.87,"total_input_tokens":950614474,"total_output_tokens":7394400,"total_runtime_s":775905.225,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.72487812208666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6e1f109d91d0740e3565","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-a9aec0c39db4ccc4c9f91cd8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2007545188144268,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1082,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[540];split=unlicensed/last-exam;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1082,"metadata":{"avgScore":0.2007545188144268,"avg_score":0.2007545188144268,"avg_score_percent":20.075451881442678,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":198.09,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":444162978,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2434975,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":4,"runs":29,"runtime":253347.01599999997,"source_flags":[],"source_harness":"kimi_code","source_row_index":540,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":198.09,"totalDurationS":253347.01599999997,"totalInputTokens":444162978,"totalOutputTokens":2434975,"total_cost_usd":198.09,"total_input_tokens":444162978,"total_output_tokens":2434975,"total_runtime_s":253347.01599999997,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.075451881442678,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-32536881b8ddb3e4c12d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a265ea2b99bc8c88db783a36","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5269304388437668,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1324,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[661];split=unlicensed/overall;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1324,"metadata":{"avgScore":0.5269304388437668,"avg_score":0.5269304388437668,"avg_score_percent":52.69304388437668,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":432.22,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":881625868,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":6896011,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":138,"runtime":637061.2620000002,"source_flags":[],"source_harness":"claude_code","source_row_index":661,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":432.22,"totalDurationS":637061.2620000002,"totalInputTokens":881625868,"totalOutputTokens":6896011,"total_cost_usd":432.22,"total_input_tokens":881625868,"total_output_tokens":6896011,"total_runtime_s":637061.2620000002,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.69304388437668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5777ad95ca616c5cba41","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a19e2231105b5be7e32da2d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5269304388437668,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":852,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[425];split=unlicensed;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":852,"metadata":{"avgScore":0.5269304388437668,"avg_score":0.5269304388437668,"avg_score_percent":52.69304388437668,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":432.22,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":881625868,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":6896011,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":138,"runtime":637061.2620000002,"source_flags":[],"source_harness":"claude_code","source_row_index":425,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":432.22,"totalDurationS":637061.2620000002,"totalInputTokens":881625868,"totalOutputTokens":6896011,"total_cost_usd":432.22,"total_input_tokens":881625868,"total_output_tokens":6896011,"total_runtime_s":637061.2620000002,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.69304388437668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-984c9ee69684a07cd12f","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-91c92cb16cdd4b1896eff549","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3297649487170496,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":594,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[296];split=licensed;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":594,"metadata":{"avgScore":0.3297649487170496,"avg_score":0.3297649487170496,"avg_score_percent":32.97649487170496,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":166.16,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":460330832,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":1067018,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":11,"runtime":99345.342,"source_flags":[],"source_harness":"kimi_code","source_row_index":296,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":166.16,"totalDurationS":99345.342,"totalInputTokens":460330832,"totalOutputTokens":1067018,"total_cost_usd":166.16,"total_input_tokens":460330832,"total_output_tokens":1067018,"total_runtime_s":99345.342,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":32.97649487170496,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5a25c925728769497634","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-91590fcf22c04a966db5d96b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.25496571532732815,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":604,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[301];split=licensed;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":604,"metadata":{"avgScore":0.25496571532732815,"avg_score":0.25496571532732815,"avg_score_percent":25.496571532732816,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.64,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":68988606,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":498389,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":11,"runtime":138843.963,"source_flags":[],"source_harness":"claude_code","source_row_index":301,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":31.64,"totalDurationS":138843.963,"totalInputTokens":68988606,"totalOutputTokens":498389,"total_cost_usd":31.64,"total_input_tokens":68988606,"total_output_tokens":498389,"total_runtime_s":138843.963,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":25.496571532732816,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-1db02eb5499f4fd1c578","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-73e6d5156fda239b91a640b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.515832318778052,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":484,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[241];split=full/overall;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":484,"metadata":{"avgScore":0.515832318778052,"avg_score":0.515832318778052,"avg_score_percent":51.583231877805204,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":605.78,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":1402173402,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":7479641,"passRate":0.28289473684210525,"pass_rate":0.28289473684210525,"pass_rate_percent":28.289473684210524,"passes":43,"runs":149,"runtime":669880.985,"source_flags":[],"source_harness":"kimi_code","source_row_index":241,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":605.78,"totalDurationS":669880.985,"totalInputTokens":1402173402,"totalOutputTokens":7479641,"total_cost_usd":605.78,"total_input_tokens":1402173402,"total_output_tokens":7479641,"total_runtime_s":669880.985,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.583231877805204,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6704399899bbd794bbcb","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-6923c7c434176ab95ce220d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.48133859388693956,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":956,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[477];split=unlicensed/full-spectrum;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":956,"metadata":{"avgScore":0.48133859388693956,"avg_score":0.48133859388693956,"avg_score_percent":48.133859388693956,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":138.27,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":283369221,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2184498,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":14,"runs":49,"runtime":160811.89200000002,"source_flags":[],"source_harness":"kimi_code","source_row_index":477,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":138.27,"totalDurationS":160811.89200000002,"totalInputTokens":283369221,"totalOutputTokens":2184498,"total_cost_usd":138.27,"total_input_tokens":283369221,"total_output_tokens":2184498,"total_runtime_s":160811.89200000002,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.133859388693956,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bff5e598b39d95e7758a","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-60aa253e2a616f881e6a7983","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5195649529905552,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":718,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[358];split=linux_only;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":718,"metadata":{"avgScore":0.5195649529905552,"avg_score":0.5195649529905552,"avg_score_percent":51.95649529905552,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":304.99,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":604383029,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":5019776,"passRate":0.2761904761904762,"pass_rate":0.2761904761904762,"pass_rate_percent":27.61904761904762,"passes":29,"runs":102,"runtime":368089.04900000006,"source_flags":[],"source_harness":"kimi_code","source_row_index":358,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":304.99,"totalDurationS":368089.04900000006,"totalInputTokens":604383029,"totalOutputTokens":5019776,"total_cost_usd":304.99,"total_input_tokens":604383029,"total_output_tokens":5019776,"total_runtime_s":368089.04900000006,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.95649529905552,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b3872f82eef660a7df29","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-5ac4cf39e7ee5b0e543a16af","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.44940530822949676,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":130,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[64];split=full/full-spectrum;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":130,"metadata":{"avgScore":0.44940530822949676,"avg_score":0.44940530822949676,"avg_score_percent":44.940530822949675,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":143.86,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":297756425,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2433240,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":54,"runtime":233738.59900000002,"source_flags":[],"source_harness":"claude_code","source_row_index":64,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":143.86,"totalDurationS":233738.59900000002,"totalInputTokens":297756425,"totalOutputTokens":2433240,"total_cost_usd":143.86,"total_input_tokens":297756425,"total_output_tokens":2433240,"total_runtime_s":233738.59900000002,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.940530822949675,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a3317ce44d6d9192a4a4","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-54fa09b11268dd8c2b2fd91d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2055276315873776,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":244,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[121];split=full/last-exam;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":244,"metadata":{"avgScore":0.2055276315873776,"avg_score":0.2055276315873776,"avg_score_percent":20.55276315873776,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":308.4,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":748185123,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":3125141,"passRate":0.10526315789473684,"pass_rate":0.10526315789473684,"pass_rate_percent":10.526315789473683,"passes":4,"runs":37,"runtime":330764.91299999994,"source_flags":[],"source_harness":"kimi_code","source_row_index":121,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":308.4,"totalDurationS":330764.91299999994,"totalInputTokens":748185123,"totalOutputTokens":3125141,"total_cost_usd":308.4,"total_input_tokens":748185123,"total_output_tokens":3125141,"total_runtime_s":330764.91299999994,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.55276315873776,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d41b803a7c7194abcb5","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-4884f8c5e7eb25cdcc397e18","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4987405398972178,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":120,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[59];split=full/full-spectrum;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":120,"metadata":{"avgScore":0.4987405398972178,"avg_score":0.4987405398972178,"avg_score_percent":49.87405398972178,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":206.05,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":465363592,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2768890,"passRate":0.2909090909090909,"pass_rate":0.2909090909090909,"pass_rate_percent":29.09090909090909,"passes":16,"runs":54,"runtime":200627.39300000004,"source_flags":[],"source_harness":"kimi_code","source_row_index":59,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":206.05,"totalDurationS":200627.39300000004,"totalInputTokens":465363592,"totalOutputTokens":2768890,"total_cost_usd":206.05,"total_input_tokens":465363592,"total_output_tokens":2768890,"total_runtime_s":200627.39300000004,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.87405398972178,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-183ac97704d7890799bf","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-42df3c93f19d2991a2738419","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5072487812208666,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":488,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[243];split=full/overall;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":488,"metadata":{"avgScore":0.5072487812208666,"avg_score":0.5072487812208666,"avg_score_percent":50.72487812208666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":463.87,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":950614474,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":7394400,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":149,"runtime":775905.225,"source_flags":[],"source_harness":"claude_code","source_row_index":243,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":463.87,"totalDurationS":775905.225,"totalInputTokens":950614474,"totalOutputTokens":7394400,"total_cost_usd":463.87,"total_input_tokens":950614474,"total_output_tokens":7394400,"total_runtime_s":775905.225,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":50.72487812208666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aba94d0f843cbed66ea3","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3725ad995e818a84a65c7794","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23142431676345085,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1078,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[538];split=unlicensed/last-exam;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1078,"metadata":{"avgScore":0.23142431676345085,"avg_score":0.23142431676345085,"avg_score_percent":23.142431676345083,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":186.26,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":410514334,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2347073,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":28,"runtime":255720.027,"source_flags":[],"source_harness":"claude_code","source_row_index":538,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":186.26,"totalDurationS":255720.027,"totalInputTokens":410514334,"totalOutputTokens":2347073,"total_cost_usd":186.26,"total_input_tokens":410514334,"total_output_tokens":2347073,"total_runtime_s":255720.027,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.142431676345083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ff2c38b63c5cc63f1217","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-2d32f66455e4a2f0505c7b76","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5303482128962862,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":850,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[424];split=unlicensed;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":850,"metadata":{"avgScore":0.5303482128962862,"avg_score":0.5303482128962862,"avg_score_percent":53.03482128962862,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.62,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":941842570,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":6412623,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":42,"runs":138,"runtime":570535.643,"source_flags":[],"source_harness":"kimi_code","source_row_index":424,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":439.62,"totalDurationS":570535.643,"totalInputTokens":941842570,"totalOutputTokens":6412623,"total_cost_usd":439.62,"total_input_tokens":941842570,"total_output_tokens":6412623,"total_runtime_s":570535.643,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.03482128962862,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c3aec33feabead91fc66","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1f03a0e3d98a8cca4494dca9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.4351458390524464,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":972,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[485];split=unlicensed/full-spectrum;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":972,"metadata":{"avgScore":0.4351458390524464,"avg_score":0.4351458390524464,"avg_score_percent":43.51458390524464,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":125.25,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":254930145,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2186567,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":49,"runtime":182599.69500000004,"source_flags":[],"source_harness":"claude_code","source_row_index":485,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":125.25,"totalDurationS":182599.69500000004,"totalInputTokens":254930145,"totalOutputTokens":2186567,"total_cost_usd":125.25,"total_input_tokens":254930145,"total_output_tokens":2186567,"total_runtime_s":182599.69500000004,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.51458390524464,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d90409a378eb64e56797","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-19f4b38fde6c4b835cf328d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.7159801555439591,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":378,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[188];split=full/near-term;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":378,"metadata":{"avgScore":0.7159801555439591,"avg_score":0.7159801555439591,"avg_score_percent":71.59801555439591,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":145.05,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":270158643,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2722015,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":239761.4640000001,"source_flags":[],"source_harness":"claude_code","source_row_index":188,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":145.05,"totalDurationS":239761.4640000001,"totalInputTokens":270158643,"totalOutputTokens":2722015,"total_cost_usd":145.05,"total_input_tokens":270158643,"total_output_tokens":2722015,"total_runtime_s":239761.4640000001,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":71.59801555439591,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a5a06f5ebe182ab13b40","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-0bcadc399c9aa05e72d39e27","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.6968586979044264,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":384,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[191];split=full/near-term;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":384,"metadata":{"avgScore":0.6968586979044264,"avg_score":0.6968586979044264,"avg_score_percent":69.68586979044264,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":129.35,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":256675448,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":2162427,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":66,"runtime":182895.38799999998,"source_flags":[],"source_harness":"kimi_code","source_row_index":191,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":129.35,"totalDurationS":182895.38799999998,"totalInputTokens":256675448,"totalOutputTokens":2162427,"total_cost_usd":129.35,"total_input_tokens":256675448,"total_output_tokens":2162427,"total_runtime_s":182895.38799999998,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":69.68586979044264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b5e49a41d487b20e0198","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-08dd0123bb050166f4f4bc6d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.5303482128962862,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d4801e71785a4e10133ae18e7035ef8b359256","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1322,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[660];split=unlicensed/overall;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1322,"metadata":{"avgScore":0.5303482128962862,"avg_score":0.5303482128962862,"avg_score_percent":53.03482128962862,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.62,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":941842570,"metric_id":"avg_score","metric_role":"secondary","model":"kimi-k3","output_tokens":6412623,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":42,"runs":138,"runtime":570535.643,"source_flags":[],"source_harness":"kimi_code","source_row_index":660,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":439.62,"totalDurationS":570535.643,"totalInputTokens":941842570,"totalOutputTokens":6412623,"total_cost_usd":439.62,"total_input_tokens":941842570,"total_output_tokens":6412623,"total_runtime_s":570535.643,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.03482128962862,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3979f51af57ec83e78b5","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-f852d65dfd2704c4b66b526a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":119,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[59];split=full/full-spectrum;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":119,"metadata":{"avgScore":0.4987405398972178,"avg_score":0.4987405398972178,"avg_score_percent":49.87405398972178,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":206.05,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":465363592,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2768890,"passRate":0.2909090909090909,"pass_rate":0.2909090909090909,"pass_rate_percent":29.09090909090909,"passes":16,"runs":54,"runtime":200627.39300000004,"source_flags":[],"source_harness":"kimi_code","source_row_index":59,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":206.05,"totalDurationS":200627.39300000004,"totalInputTokens":465363592,"totalOutputTokens":2768890,"total_cost_usd":206.05,"total_input_tokens":465363592,"total_output_tokens":2768890,"total_runtime_s":200627.39300000004,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.09090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f625b72657bd8a339e18","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f663dddfaf9cd1a5b6827234","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2761904761904762,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":713,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[356];split=linux_only;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":713,"metadata":{"avgScore":0.5278637305133695,"avg_score":0.5278637305133695,"avg_score_percent":52.78637305133695,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":251.07,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":465187955,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":4686442,"passRate":0.2761904761904762,"pass_rate":0.2761904761904762,"pass_rate_percent":27.61904761904762,"passes":29,"runs":102,"runtime":393848.4819999999,"source_flags":[],"source_harness":"claude_code","source_row_index":356,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":251.07,"totalDurationS":393848.4819999999,"totalInputTokens":465187955,"totalOutputTokens":4686442,"total_cost_usd":251.07,"total_input_tokens":465187955,"total_output_tokens":4686442,"total_runtime_s":393848.4819999999,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.61904761904762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7dcfe54789b1eced04f5","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f65517744077ad7aa69c26c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1077,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[538];split=unlicensed/last-exam;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1077,"metadata":{"avgScore":0.23142431676345085,"avg_score":0.23142431676345085,"avg_score_percent":23.142431676345083,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":186.26,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":410514334,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2347073,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":3,"runs":28,"runtime":255720.027,"source_flags":[],"source_harness":"claude_code","source_row_index":538,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":186.26,"totalDurationS":255720.027,"totalInputTokens":410514334,"totalOutputTokens":2347073,"total_cost_usd":186.26,"total_input_tokens":410514334,"total_output_tokens":2347073,"total_runtime_s":255720.027,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-95a971395bd7e5706787","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e9ebdc8fad2b98569db20653","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28368794326241137,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1323,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[661];split=unlicensed/overall;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1323,"metadata":{"avgScore":0.5269304388437668,"avg_score":0.5269304388437668,"avg_score_percent":52.69304388437668,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":432.22,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":881625868,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":6896011,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":138,"runtime":637061.2620000002,"source_flags":[],"source_harness":"claude_code","source_row_index":661,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":432.22,"totalDurationS":637061.2620000002,"totalInputTokens":881625868,"totalOutputTokens":6896011,"total_cost_usd":432.22,"total_input_tokens":881625868,"total_output_tokens":6896011,"total_runtime_s":637061.2620000002,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.368794326241137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-93fbb25513075e29631c","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-e882ec4a0551d8c919f20176","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.10526315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":243,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[121];split=full/last-exam;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":243,"metadata":{"avgScore":0.2055276315873776,"avg_score":0.2055276315873776,"avg_score_percent":20.55276315873776,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":308.4,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":748185123,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":3125141,"passRate":0.10526315789473684,"pass_rate":0.10526315789473684,"pass_rate_percent":10.526315789473683,"passes":4,"runs":37,"runtime":330764.91299999994,"source_flags":[],"source_harness":"kimi_code","source_row_index":121,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":37,"tokens":null,"totalCostUsd":308.4,"totalDurationS":330764.91299999994,"totalInputTokens":748185123,"totalOutputTokens":3125141,"total_cost_usd":308.4,"total_input_tokens":748185123,"total_output_tokens":3125141,"total_runtime_s":330764.91299999994,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.526315789473683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-05e7e1f18a7a0ec15858","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-dde400cbed85efcfa88b312b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28368794326241137,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":851,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[425];split=unlicensed;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":851,"metadata":{"avgScore":0.5269304388437668,"avg_score":0.5269304388437668,"avg_score_percent":52.69304388437668,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":432.22,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":881625868,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":6896011,"passRate":0.28368794326241137,"pass_rate":0.28368794326241137,"pass_rate_percent":28.368794326241137,"passes":40,"runs":138,"runtime":637061.2620000002,"source_flags":[],"source_harness":"claude_code","source_row_index":425,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":432.22,"totalDurationS":637061.2620000002,"totalInputTokens":881625868,"totalOutputTokens":6896011,"total_cost_usd":432.22,"total_input_tokens":881625868,"total_output_tokens":6896011,"total_runtime_s":637061.2620000002,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.368794326241137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f03089f2434ce79f3964","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c57a99ba4ea98784b5fd2d30","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40298507462686567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":377,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[188];split=full/near-term;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":377,"metadata":{"avgScore":0.7159801555439591,"avg_score":0.7159801555439591,"avg_score_percent":71.59801555439591,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":145.05,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":270158643,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2722015,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":239761.4640000001,"source_flags":[],"source_harness":"claude_code","source_row_index":188,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":145.05,"totalDurationS":239761.4640000001,"totalInputTokens":270158643,"totalOutputTokens":2722015,"total_cost_usd":145.05,"total_input_tokens":270158643,"total_output_tokens":2722015,"total_runtime_s":239761.4640000001,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.298507462686565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dee7d0ff5cf79f05ef0e","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-b70e301d425fab40c859bf2b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28289473684210525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":483,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[241];split=full/overall;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":483,"metadata":{"avgScore":0.515832318778052,"avg_score":0.515832318778052,"avg_score_percent":51.583231877805204,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":605.78,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":1402173402,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":7479641,"passRate":0.28289473684210525,"pass_rate":0.28289473684210525,"pass_rate_percent":28.289473684210524,"passes":43,"runs":149,"runtime":669880.985,"source_flags":[],"source_harness":"kimi_code","source_row_index":241,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":605.78,"totalDurationS":669880.985,"totalInputTokens":1402173402,"totalOutputTokens":7479641,"total_cost_usd":605.78,"total_input_tokens":1402173402,"total_output_tokens":7479641,"total_runtime_s":669880.985,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.289473684210524,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6e6a2445e26628c415e2","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-b4ad6c25991c111be8c53161","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.13333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1081,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[540];split=unlicensed/last-exam;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1081,"metadata":{"avgScore":0.2007545188144268,"avg_score":0.2007545188144268,"avg_score_percent":20.075451881442678,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":198.09,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":444162978,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2434975,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":4,"runs":29,"runtime":253347.01599999997,"source_flags":[],"source_harness":"kimi_code","source_row_index":540,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":198.09,"totalDurationS":253347.01599999997,"totalInputTokens":444162978,"totalOutputTokens":2434975,"total_cost_usd":198.09,"total_input_tokens":444162978,"total_output_tokens":2434975,"total_runtime_s":253347.01599999997,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.333333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a2d4efba3e1be2605e30","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-8e78e28d908b4c43c17d79b6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2978723404255319,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1321,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[660];split=unlicensed/overall;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1321,"metadata":{"avgScore":0.5303482128962862,"avg_score":0.5303482128962862,"avg_score_percent":53.03482128962862,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.62,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":941842570,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":6412623,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":42,"runs":138,"runtime":570535.643,"source_flags":[],"source_harness":"kimi_code","source_row_index":660,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":439.62,"totalDurationS":570535.643,"totalInputTokens":941842570,"totalOutputTokens":6412623,"total_cost_usd":439.62,"total_input_tokens":941842570,"total_output_tokens":6412623,"total_runtime_s":570535.643,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.78723404255319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2add7043b34e19e6a170","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-8df5ea7eb4855dbef91b6c67","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1219,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[609];split=unlicensed/near-term;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1219,"metadata":{"avgScore":0.6968586979044264,"avg_score":0.6968586979044264,"avg_score_percent":69.68586979044264,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":129.35,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":256675448,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2162427,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":66,"runtime":182895.38799999998,"source_flags":[],"source_harness":"kimi_code","source_row_index":609,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":129.35,"totalDurationS":182895.38799999998,"totalInputTokens":256675448,"totalOutputTokens":2162427,"total_cost_usd":129.35,"total_input_tokens":256675448,"total_output_tokens":2162427,"total_runtime_s":182895.38799999998,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dd24803775fb156e402b","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-7bb5759e62daf480e4721714","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2978723404255319,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":849,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[424];split=unlicensed;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":849,"metadata":{"avgScore":0.5303482128962862,"avg_score":0.5303482128962862,"avg_score_percent":53.03482128962862,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":439.62,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":941842570,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":6412623,"passRate":0.2978723404255319,"pass_rate":0.2978723404255319,"pass_rate_percent":29.78723404255319,"passes":42,"runs":138,"runtime":570535.643,"source_flags":[],"source_harness":"kimi_code","source_row_index":424,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":439.62,"totalDurationS":570535.643,"totalInputTokens":941842570,"totalOutputTokens":6412623,"total_cost_usd":439.62,"total_input_tokens":941842570,"total_output_tokens":6412623,"total_runtime_s":570535.643,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.78723404255319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-156f10e7c738c9c9c6cc","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7a70478c3fa720a14a351f04","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.07894736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":245,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[122];split=full/last-exam;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":245,"metadata":{"avgScore":0.2049303255658983,"avg_score":0.2049303255658983,"avg_score_percent":20.49303255658983,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":200.05,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":437612423,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2624214,"passRate":0.07894736842105263,"pass_rate":0.07894736842105263,"pass_rate_percent":7.894736842105263,"passes":3,"runs":36,"runtime":363134.68600000005,"source_flags":[],"source_harness":"claude_code","source_row_index":122,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":200.05,"totalDurationS":363134.68600000005,"totalInputTokens":437612423,"totalOutputTokens":2624214,"total_cost_usd":200.05,"total_input_tokens":437612423,"total_output_tokens":2624214,"total_runtime_s":363134.68600000005,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":7.894736842105263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-854b230d3a1bcac6ba51","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-65c3e1453c6668ba43a69cc8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":955,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[477];split=unlicensed/full-spectrum;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":955,"metadata":{"avgScore":0.48133859388693956,"avg_score":0.48133859388693956,"avg_score_percent":48.133859388693956,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":138.27,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":283369221,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2184498,"passRate":0.28,"pass_rate":0.28,"pass_rate_percent":28.000000000000004,"passes":14,"runs":49,"runtime":160811.89200000002,"source_flags":[],"source_harness":"kimi_code","source_row_index":477,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":138.27,"totalDurationS":160811.89200000002,"totalInputTokens":283369221,"totalOutputTokens":2184498,"total_cost_usd":138.27,"total_input_tokens":283369221,"total_output_tokens":2184498,"total_runtime_s":160811.89200000002,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f36d3e20b6ae94d083e8","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-59eaf3f1fea483fb0ded9696","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.23636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":129,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[64];split=full/full-spectrum;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":129,"metadata":{"avgScore":0.44940530822949676,"avg_score":0.44940530822949676,"avg_score_percent":44.940530822949675,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":143.86,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":297756425,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2433240,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":54,"runtime":233738.59900000002,"source_flags":[],"source_harness":"claude_code","source_row_index":64,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":143.86,"totalDurationS":233738.59900000002,"totalInputTokens":297756425,"totalOutputTokens":2433240,"total_cost_usd":143.86,"total_input_tokens":297756425,"total_output_tokens":2433240,"total_runtime_s":233738.59900000002,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.636363636363637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-102443cc949c1c04d9b8","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-5837a6e88b8b4a47f704a459","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":593,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[296];split=licensed;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":593,"metadata":{"avgScore":0.3297649487170496,"avg_score":0.3297649487170496,"avg_score_percent":32.97649487170496,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":166.16,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":460330832,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":1067018,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":11,"runtime":99345.342,"source_flags":[],"source_harness":"kimi_code","source_row_index":296,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":166.16,"totalDurationS":99345.342,"totalInputTokens":460330832,"totalOutputTokens":1067018,"total_cost_usd":166.16,"total_input_tokens":460330832,"total_output_tokens":1067018,"total_runtime_s":99345.342,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d6fdcc6860f628a81298","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-554de4d627ed1c3aeda9f6b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.2761904761904762,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":717,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[358];split=linux_only;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":717,"metadata":{"avgScore":0.5195649529905552,"avg_score":0.5195649529905552,"avg_score_percent":51.95649529905552,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":304.99,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":604383029,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":5019776,"passRate":0.2761904761904762,"pass_rate":0.2761904761904762,"pass_rate_percent":27.61904761904762,"passes":29,"runs":102,"runtime":368089.04900000006,"source_flags":[],"source_harness":"kimi_code","source_row_index":358,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":304.99,"totalDurationS":368089.04900000006,"totalInputTokens":604383029,"totalOutputTokens":5019776,"total_cost_usd":304.99,"total_input_tokens":604383029,"total_output_tokens":5019776,"total_runtime_s":368089.04900000006,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":27.61904761904762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a05d1f9872deebea062d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-4c0179ad9ca11bbc91ef708c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":383,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[191];split=full/near-term;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":383,"metadata":{"avgScore":0.6968586979044264,"avg_score":0.6968586979044264,"avg_score_percent":69.68586979044264,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":129.35,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":256675448,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2162427,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":66,"runtime":182895.38799999998,"source_flags":[],"source_harness":"kimi_code","source_row_index":191,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":129.35,"totalDurationS":182895.38799999998,"totalInputTokens":256675448,"totalOutputTokens":2162427,"total_cost_usd":129.35,"total_input_tokens":256675448,"total_output_tokens":2162427,"total_runtime_s":182895.38799999998,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a5a2545d09b5907bc381","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"kimi-code","harnessId":"kimi-code","httpStatus":200,"id":"pub-4638d257d7d5a26f9bec0e2d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"kimi-code","harness_variant":"thinking-max","source_harness":"kimi_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.28289473684210525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":11,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[5];split=full;harness=kimi_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":11,"metadata":{"avgScore":0.515832318778052,"avg_score":0.515832318778052,"avg_score_percent":51.583231877805204,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":605.78,"costSource":"db","cost_source":"db","harness":"kimi_code","harness_id":"kimi-code","harness_variant":"thinking-max","input_tokens":1402173402,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":7479641,"passRate":0.28289473684210525,"pass_rate":0.28289473684210525,"pass_rate_percent":28.289473684210524,"passes":43,"runs":149,"runtime":669880.985,"source_flags":[],"source_harness":"kimi_code","source_row_index":5,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":605.78,"totalDurationS":669880.985,"totalInputTokens":1402173402,"totalOutputTokens":7479641,"total_cost_usd":605.78,"total_input_tokens":1402173402,"total_output_tokens":7479641,"total_runtime_s":669880.985,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.289473684210524,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cd88a5f8b8d7f4e0f0bd","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-456526ad7d409815ce604307","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":487,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[243];split=full/overall;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":487,"metadata":{"avgScore":0.5072487812208666,"avg_score":0.5072487812208666,"avg_score_percent":50.72487812208666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":463.87,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":950614474,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":7394400,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":149,"runtime":775905.225,"source_flags":[],"source_harness":"claude_code","source_row_index":243,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":463.87,"totalDurationS":775905.225,"totalInputTokens":950614474,"totalOutputTokens":7394400,"total_cost_usd":463.87,"total_input_tokens":950614474,"total_output_tokens":7394400,"total_runtime_s":775905.225,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-037b0aa696ef5b643307","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3c29fb9fe05ab1bd76803961","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":15,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[7];split=full;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":15,"metadata":{"avgScore":0.5072487812208666,"avg_score":0.5072487812208666,"avg_score_percent":50.72487812208666,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":463.87,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":950614474,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":7394400,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":149,"runtime":775905.225,"source_flags":[],"source_harness":"claude_code","source_row_index":7,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":149,"tokens":null,"totalCostUsd":463.87,"totalDurationS":775905.225,"totalInputTokens":950614474,"totalOutputTokens":7394400,"total_cost_usd":463.87,"total_input_tokens":950614474,"total_output_tokens":7394400,"total_runtime_s":775905.225,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c93873bac1af7ce090fd","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2e7938d6cc89333e1c475b5a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.40298507462686567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1213,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[606];split=unlicensed/near-term;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1213,"metadata":{"avgScore":0.7159801555439591,"avg_score":0.7159801555439591,"avg_score_percent":71.59801555439591,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":145.05,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":270158643,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2722015,"passRate":0.40298507462686567,"pass_rate":0.40298507462686567,"pass_rate_percent":40.298507462686565,"passes":27,"runs":67,"runtime":239761.4640000001,"source_flags":[],"source_harness":"claude_code","source_row_index":606,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":145.05,"totalDurationS":239761.4640000001,"totalInputTokens":270158643,"totalOutputTokens":2722015,"total_cost_usd":145.05,"total_input_tokens":270158643,"total_output_tokens":2722015,"total_runtime_s":239761.4640000001,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.298507462686565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-02dc86c9683a8203b039","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-288bc7c091c7730f9195c161","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":603,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[301];split=licensed;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":603,"metadata":{"avgScore":0.25496571532732815,"avg_score":0.25496571532732815,"avg_score_percent":25.496571532732816,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.64,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":68988606,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":498389,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":1,"runs":11,"runtime":138843.963,"source_flags":[],"source_harness":"claude_code","source_row_index":301,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":11,"tokens":null,"totalCostUsd":31.64,"totalDurationS":138843.963,"totalInputTokens":68988606,"totalOutputTokens":498389,"total_cost_usd":31.64,"total_input_tokens":68988606,"total_output_tokens":498389,"total_runtime_s":138843.963,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-66613a9c825a3f48831a","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1b38e767bc7c18c65ed8ceb3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","missing_source_date","split_specific"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dffc7c16e74f7edf3839b620dd2a9332ee1e722","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":971,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[485];split=unlicensed/full-spectrum;harness=claude_code;model=kimi-k3;variant=thinking-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":971,"metadata":{"avgScore":0.4351458390524464,"avg_score":0.4351458390524464,"avg_score_percent":43.51458390524464,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":125.25,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-max","input_tokens":254930145,"metric_id":"pass_rate","metric_role":"primary","model":"kimi-k3","output_tokens":2186567,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":49,"runtime":182599.69500000004,"source_flags":[],"source_harness":"claude_code","source_row_index":485,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":125.25,"totalDurationS":182599.69500000004,"totalInputTokens":254930145,"totalOutputTokens":2186567,"total_cost_usd":125.25,"total_input_tokens":254930145,"total_output_tokens":2186567,"total_runtime_s":182599.69500000004,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-da8ffa723a3880677a0e","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e5314f14194590dff4f969b2","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.21744599999999994,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1058,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[528];split=unlicensed/full-spectrum;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1058,"metadata":{"avgScore":0.21744599999999994,"avg_score":0.21744599999999994,"avg_score_percent":21.744599999999995,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":13.46,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":125903257.16666667,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":1210919.8333333333,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":9,"runs":90,"runtime":99113.22360567635,"source_flags":[],"source_harness":"openclaw","source_row_index":528,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":13.46,"totalDurationS":99113.22360567635,"totalInputTokens":125903257.16666667,"totalOutputTokens":1210919.8333333333,"total_cost_usd":13.46,"total_input_tokens":125903257.16666667,"total_output_tokens":1210919.8333333333,"total_runtime_s":99113.22360567635,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":21.744599999999995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b0149201d2950d6f4abc","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d644e545a04ba2c50824bd53","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23555956574853623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":578,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[288];split=full/overall;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":578,"metadata":{"avgScore":0.23555956574853623,"avg_score":0.23555956574853623,"avg_score_percent":23.555956574853624,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.34,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":394850494.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":4001432.3333333335,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":261,"runtime":386539.2838370383,"source_flags":[],"source_harness":"openclaw","source_row_index":288,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":43.34,"totalDurationS":386539.2838370383,"totalInputTokens":394850494.1666667,"totalOutputTokens":4001432.3333333335,"total_cost_usd":43.34,"total_input_tokens":394850494.1666667,"total_output_tokens":4001432.3333333335,"total_runtime_s":386539.2838370383,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.555956574853624,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-aca2e39261db9cb4b7ee","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d0db3037301306f20b9590b8","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26465217947397845,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":822,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[410];split=linux_only;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":822,"metadata":{"avgScore":0.26465217947397845,"avg_score":0.26465217947397845,"avg_score_percent":26.465217947397846,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.54,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":286441573.5,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":3005020.6666666665,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":20,"runs":197,"runtime":185037.9774775899,"source_flags":[],"source_harness":"openclaw","source_row_index":410,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":31.54,"totalDurationS":185037.9774775899,"totalInputTokens":286441573.5,"totalOutputTokens":3005020.6666666665,"total_cost_usd":31.54,"total_input_tokens":286441573.5,"total_output_tokens":3005020.6666666665,"total_runtime_s":185037.9774775899,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.465217947397846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e45987928b62ea5f135b","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cac4f2f20d998f6d766fda00","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.23555956574853623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":106,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[52];split=full;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":106,"metadata":{"avgScore":0.23555956574853623,"avg_score":0.23555956574853623,"avg_score_percent":23.555956574853624,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.34,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":394850494.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":4001432.3333333335,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":261,"runtime":386539.2838370383,"source_flags":[],"source_harness":"openclaw","source_row_index":52,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":43.34,"totalDurationS":386539.2838370383,"totalInputTokens":394850494.1666667,"totalOutputTokens":4001432.3333333335,"total_cost_usd":43.34,"total_input_tokens":394850494.1666667,"total_output_tokens":4001432.3333333335,"total_runtime_s":386539.2838370383,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":23.555956574853624,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-70872e7af08d01e795dc","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bce9ad0f54243ce2ecc113df","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3508749428062655,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1298,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[648];split=unlicensed/near-term;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1298,"metadata":{"avgScore":0.3508749428062655,"avg_score":0.3508749428062655,"avg_score_percent":35.08749428062655,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.77,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":105443753,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":1528714,"passRate":0.11940298507462686,"pass_rate":0.11940298507462686,"pass_rate_percent":11.940298507462686,"passes":16,"runs":123,"runtime":157005.70431608308,"source_flags":[],"source_harness":"openclaw","source_row_index":648,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":12.77,"totalDurationS":157005.70431608308,"totalInputTokens":105443753,"totalOutputTokens":1528714,"total_cost_usd":12.77,"total_input_tokens":105443753,"total_output_tokens":1528714,"total_runtime_s":157005.70431608308,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.08749428062655,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d7bd6c625f3cd2d492e7","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a8eb1a95304a12d16b893042","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.20754532410468568,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":222,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[110];split=full/full-spectrum;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":222,"metadata":{"avgScore":0.20754532410468568,"avg_score":0.20754532410468568,"avg_score_percent":20.754532410468567,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":16.57,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":159897475.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":1362370.8333333333,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":9,"runs":98,"runtime":125442.30073191645,"source_flags":[],"source_harness":"openclaw","source_row_index":110,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":16.57,"totalDurationS":125442.30073191645,"totalInputTokens":159897475.16666666,"totalOutputTokens":1362370.8333333333,"total_cost_usd":16.57,"total_input_tokens":159897475.16666666,"total_output_tokens":1362370.8333333333,"total_runtime_s":125442.30073191645,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":20.754532410468567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b9d49794e2a8fb65282f","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-82013fe23c1135f344570aa7","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2500876678582964,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1414,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[706];split=unlicensed/overall;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1414,"metadata":{"avgScore":0.2500876678582964,"avg_score":0.2500876678582964,"avg_score_percent":25.00876678582964,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.49,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":342350334.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":3820122.8333333335,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":251,"runtime":356215.3619929354,"source_flags":[],"source_harness":"openclaw","source_row_index":706,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":38.49,"totalDurationS":356215.3619929354,"totalInputTokens":342350334.1666667,"totalOutputTokens":3820122.8333333335,"total_cost_usd":38.49,"total_input_tokens":342350334.1666667,"total_output_tokens":3820122.8333333335,"total_runtime_s":356215.3619929354,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":25.00876678582964,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ac39772c616f8f31e6cb","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7b33cb2365a9e8a153fe66b9","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2500876678582964,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":942,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[470];split=unlicensed;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":942,"metadata":{"avgScore":0.2500876678582964,"avg_score":0.2500876678582964,"avg_score_percent":25.00876678582964,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.49,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":342350334.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":3820122.8333333335,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":251,"runtime":356215.3619929354,"source_flags":[],"source_harness":"openclaw","source_row_index":470,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":38.49,"totalDurationS":356215.3619929354,"totalInputTokens":342350334.1666667,"totalOutputTokens":3820122.8333333335,"total_cost_usd":38.49,"total_input_tokens":342350334.1666667,"total_output_tokens":3820122.8333333335,"total_runtime_s":356215.3619929354,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":25.00876678582964,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cc9d95fbb16c9b25bc44","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6ab1bcc2940cc1d55a1d5749","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.04933571143251942,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":692,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[345];split=licensed;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":692,"metadata":{"avgScore":0.04933571143251942,"avg_score":0.04933571143251942,"avg_score_percent":4.9335711432519425,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":52500160,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":181309.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":30323.92184410304,"source_flags":[],"source_harness":"openclaw","source_row_index":345,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":4.85,"totalDurationS":30323.92184410304,"totalInputTokens":52500160,"totalOutputTokens":181309.5,"total_cost_usd":4.85,"total_input_tokens":52500160,"total_output_tokens":181309.5,"total_runtime_s":30323.92184410304,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.9335711432519425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-09c1fb01464f2612e11e","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5896b8d3071b45d0952d8395","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3508749428062655,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":462,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[230];split=full/near-term;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":462,"metadata":{"avgScore":0.3508749428062655,"avg_score":0.3508749428062655,"avg_score_percent":35.08749428062655,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.77,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":105443753,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":1528714,"passRate":0.11940298507462686,"pass_rate":0.11940298507462686,"pass_rate_percent":11.940298507462686,"passes":16,"runs":123,"runtime":157005.70431608308,"source_flags":[],"source_harness":"openclaw","source_row_index":230,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":12.77,"totalDurationS":157005.70431608308,"totalInputTokens":105443753,"totalOutputTokens":1528714,"total_cost_usd":12.77,"total_input_tokens":105443753,"total_output_tokens":1528714,"total_runtime_s":157005.70431608308,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":35.08749428062655,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7df66e4c266e3e354fec","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-335d7931794c469ad0e73a46","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.043168,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1170,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[584];split=unlicensed/last-exam;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1170,"metadata":{"avgScore":0.043168,"avg_score":0.043168,"avg_score_percent":4.3168,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":13.02,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":116658991.5,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":1200145,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":50,"runtime":113242.63516030132,"source_flags":[],"source_harness":"openclaw","source_row_index":584,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":13.02,"totalDurationS":113242.63516030132,"totalInputTokens":116658991.5,"totalOutputTokens":1200145,"total_cost_usd":13.02,"total_input_tokens":116658991.5,"total_output_tokens":1200145,"total_runtime_s":113242.63516030132,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.3168,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9cc2d3a802b662c5973a","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0743421a9f6679f3f8cb065b","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03408,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140da44bb8fb7b92bd2d60af9ba3f6c8e17e0e76","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":336,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[167];split=full/last-exam;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":336,"metadata":{"avgScore":0.03408,"avg_score":0.03408,"avg_score_percent":3.408,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":14.92,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":135802977,"metric_id":"avg_score","metric_role":"secondary","model":"mimo-v2-5","output_tokens":1277989.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":55,"runtime":118979.19944043171,"source_flags":[],"source_harness":"openclaw","source_row_index":167,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":14.92,"totalDurationS":118979.19944043171,"totalInputTokens":135802977,"totalOutputTokens":1277989.5,"total_cost_usd":14.92,"total_input_tokens":135802977,"total_output_tokens":1277989.5,"total_runtime_s":118979.19944043171,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.408,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-810f231c78fd4c6d685b","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-fe767db30216600765f74a76","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08552631578947369,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":577,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[288];split=full/overall;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":577,"metadata":{"avgScore":0.23555956574853623,"avg_score":0.23555956574853623,"avg_score_percent":23.555956574853624,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.34,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":394850494.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":4001432.3333333335,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":261,"runtime":386539.2838370383,"source_flags":[],"source_harness":"openclaw","source_row_index":288,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":43.34,"totalDurationS":386539.2838370383,"totalInputTokens":394850494.1666667,"totalOutputTokens":4001432.3333333335,"total_cost_usd":43.34,"total_input_tokens":394850494.1666667,"total_output_tokens":4001432.3333333335,"total_runtime_s":386539.2838370383,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.552631578947368,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f86562fbce59072a13f0","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e615a8ce7160d4cca31e377e","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09219858156028368,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1413,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[706];split=unlicensed/overall;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1413,"metadata":{"avgScore":0.2500876678582964,"avg_score":0.2500876678582964,"avg_score_percent":25.00876678582964,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.49,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":342350334.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":3820122.8333333335,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":251,"runtime":356215.3619929354,"source_flags":[],"source_harness":"openclaw","source_row_index":706,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":38.49,"totalDurationS":356215.3619929354,"totalInputTokens":342350334.1666667,"totalOutputTokens":3820122.8333333335,"total_cost_usd":38.49,"total_input_tokens":342350334.1666667,"total_output_tokens":3820122.8333333335,"total_runtime_s":356215.3619929354,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.219858156028367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e7e8726b971399ff2460","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c16fef91bc306edc131f2840","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":335,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[167];split=full/last-exam;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":335,"metadata":{"avgScore":0.03408,"avg_score":0.03408,"avg_score_percent":3.408,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":14.92,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":135802977,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":1277989.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":55,"runtime":118979.19944043171,"source_flags":[],"source_harness":"openclaw","source_row_index":167,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":14.92,"totalDurationS":118979.19944043171,"totalInputTokens":135802977,"totalOutputTokens":1277989.5,"total_cost_usd":14.92,"total_input_tokens":135802977,"total_output_tokens":1277989.5,"total_runtime_s":118979.19944043171,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9100cdc84d9fdaafdcf9","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a96e5be366632d0a05effe10","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09090909090909091,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":221,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[110];split=full/full-spectrum;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":221,"metadata":{"avgScore":0.20754532410468568,"avg_score":0.20754532410468568,"avg_score_percent":20.754532410468567,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":16.57,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":159897475.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":1362370.8333333333,"passRate":0.09090909090909091,"pass_rate":0.09090909090909091,"pass_rate_percent":9.090909090909092,"passes":9,"runs":98,"runtime":125442.30073191645,"source_flags":[],"source_harness":"openclaw","source_row_index":110,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":16.57,"totalDurationS":125442.30073191645,"totalInputTokens":159897475.16666666,"totalOutputTokens":1362370.8333333333,"total_cost_usd":16.57,"total_input_tokens":159897475.16666666,"total_output_tokens":1362370.8333333333,"total_runtime_s":125442.30073191645,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.090909090909092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b4914e288de530bb5cea","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9009df64116cc2eadd50e10f","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09219858156028368,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":941,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[470];split=unlicensed;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":941,"metadata":{"avgScore":0.2500876678582964,"avg_score":0.2500876678582964,"avg_score_percent":25.00876678582964,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":38.49,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":342350334.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":3820122.8333333335,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":251,"runtime":356215.3619929354,"source_flags":[],"source_harness":"openclaw","source_row_index":470,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":38.49,"totalDurationS":356215.3619929354,"totalInputTokens":342350334.1666667,"totalOutputTokens":3820122.8333333335,"total_cost_usd":38.49,"total_input_tokens":342350334.1666667,"total_output_tokens":3820122.8333333335,"total_runtime_s":356215.3619929354,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.219858156028367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0ececc960bc1a59795b9","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8f93c4187d8ef03d90f3baea","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":821,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[410];split=linux_only;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":821,"metadata":{"avgScore":0.26465217947397845,"avg_score":0.26465217947397845,"avg_score_percent":26.465217947397846,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":31.54,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":286441573.5,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":3005020.6666666665,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":20,"runs":197,"runtime":185037.9774775899,"source_flags":[],"source_harness":"openclaw","source_row_index":410,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":31.54,"totalDurationS":185037.9774775899,"totalInputTokens":286441573.5,"totalOutputTokens":3005020.6666666665,"total_cost_usd":31.54,"total_input_tokens":286441573.5,"total_output_tokens":3005020.6666666665,"total_runtime_s":185037.9774775899,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-b75af523e937e8b3851b","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8d251286ad9d76bb4832f2c1","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08552631578947369,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":105,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[52];split=full;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":105,"metadata":{"avgScore":0.23555956574853623,"avg_score":0.23555956574853623,"avg_score_percent":23.555956574853624,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":43.34,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":394850494.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":4001432.3333333335,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":261,"runtime":386539.2838370383,"source_flags":[],"source_harness":"openclaw","source_row_index":52,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":43.34,"totalDurationS":386539.2838370383,"totalInputTokens":394850494.1666667,"totalOutputTokens":4001432.3333333335,"total_cost_usd":43.34,"total_input_tokens":394850494.1666667,"total_output_tokens":4001432.3333333335,"total_runtime_s":386539.2838370383,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.552631578947368,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5117c2d9d36ea1957359","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6085b66d68eaa719aa9ad7e6","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":691,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[345];split=licensed;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":691,"metadata":{"avgScore":0.04933571143251942,"avg_score":0.04933571143251942,"avg_score_percent":4.9335711432519425,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":4.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":52500160,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":181309.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":30323.92184410304,"source_flags":[],"source_harness":"openclaw","source_row_index":345,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":4.85,"totalDurationS":30323.92184410304,"totalInputTokens":52500160,"totalOutputTokens":181309.5,"total_cost_usd":4.85,"total_input_tokens":52500160,"total_output_tokens":181309.5,"total_runtime_s":30323.92184410304,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-43d9afd9e75224ec4a61","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-55834ba24b1e0c435ec08605","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11940298507462686,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1297,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[648];split=unlicensed/near-term;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1297,"metadata":{"avgScore":0.3508749428062655,"avg_score":0.3508749428062655,"avg_score_percent":35.08749428062655,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.77,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":105443753,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":1528714,"passRate":0.11940298507462686,"pass_rate":0.11940298507462686,"pass_rate_percent":11.940298507462686,"passes":16,"runs":123,"runtime":157005.70431608308,"source_flags":[],"source_harness":"openclaw","source_row_index":648,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":12.77,"totalDurationS":157005.70431608308,"totalInputTokens":105443753,"totalOutputTokens":1528714,"total_cost_usd":12.77,"total_input_tokens":105443753,"total_output_tokens":1528714,"total_runtime_s":157005.70431608308,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.940298507462686,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-04cf15ddf7de3fda82d4","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-32084d3584bc933f585dba06","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11940298507462686,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":461,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[230];split=full/near-term;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":461,"metadata":{"avgScore":0.3508749428062655,"avg_score":0.3508749428062655,"avg_score_percent":35.08749428062655,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":12.77,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":105443753,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":1528714,"passRate":0.11940298507462686,"pass_rate":0.11940298507462686,"pass_rate_percent":11.940298507462686,"passes":16,"runs":123,"runtime":157005.70431608308,"source_flags":[],"source_harness":"openclaw","source_row_index":230,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":12.77,"totalDurationS":157005.70431608308,"totalInputTokens":105443753,"totalOutputTokens":1528714,"total_cost_usd":12.77,"total_input_tokens":105443753,"total_output_tokens":1528714,"total_runtime_s":157005.70431608308,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.940298507462686,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d5e30db4dd77037f7a90","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-279119b4f8a0fe35f3c7186b","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1057,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[528];split=unlicensed/full-spectrum;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1057,"metadata":{"avgScore":0.21744599999999994,"avg_score":0.21744599999999994,"avg_score_percent":21.744599999999995,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":13.46,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":125903257.16666667,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":1210919.8333333333,"passRate":0.1,"pass_rate":0.1,"pass_rate_percent":10.0,"passes":9,"runs":90,"runtime":99113.22360567635,"source_flags":[],"source_harness":"openclaw","source_row_index":528,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":13.46,"totalDurationS":99113.22360567635,"totalInputTokens":125903257.16666667,"totalOutputTokens":1210919.8333333333,"total_cost_usd":13.46,"total_input_tokens":125903257.16666667,"total_output_tokens":1210919.8333333333,"total_runtime_s":99113.22360567635,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3994ff65fd52f30953af","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1215e6598f90c7815bb72027","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiMo-V2.5","modelRef":"mimo-v2-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07e43924c05207f84accb9b35f9fdaaf48037d53","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1169,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[584];split=unlicensed/last-exam;harness=openclaw;model=mimo-v2-5;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1169,"metadata":{"avgScore":0.043168,"avg_score":0.043168,"avg_score_percent":4.3168,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":13.02,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":116658991.5,"metric_id":"pass_rate","metric_role":"primary","model":"mimo-v2-5","output_tokens":1200145,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":50,"runtime":113242.63516030132,"source_flags":[],"source_harness":"openclaw","source_row_index":584,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":13.02,"totalDurationS":113242.63516030132,"totalInputTokens":116658991.5,"totalOutputTokens":1200145,"total_cost_usd":13.02,"total_input_tokens":116658991.5,"total_output_tokens":1200145,"total_runtime_s":113242.63516030132,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4eb4b21aceedfcbca74b","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f0cbbd9fe3f9fc4c46e4d5a9","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.016203333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1186,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[592];split=unlicensed/last-exam;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1186,"metadata":{"avgScore":0.016203333333333333,"avg_score":0.016203333333333333,"avg_score_percent":1.6203333333333334,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.37,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86177572.5,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":1194918.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":43,"runtime":109949.42864178025,"source_flags":[],"source_harness":"openclaw","source_row_index":592,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":8.37,"totalDurationS":109949.42864178025,"totalInputTokens":86177572.5,"totalOutputTokens":1194918.8333333333,"total_cost_usd":8.37,"total_input_tokens":86177572.5,"total_output_tokens":1194918.8333333333,"total_runtime_s":109949.42864178025,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":1.6203333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1a24fecff639257f691","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dce56a2d22221ce6ca960abd","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24547984743567736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1302,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[650];split=unlicensed/near-term;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1302,"metadata":{"avgScore":0.24547984743567736,"avg_score":0.24547984743567736,"avg_score_percent":24.547984743567735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.86,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98237249.5,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":1413585,"passRate":0.1044776119402985,"pass_rate":0.1044776119402985,"pass_rate_percent":10.44776119402985,"passes":12,"runs":105,"runtime":228093.81403760775,"source_flags":[],"source_harness":"openclaw","source_row_index":650,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":8.86,"totalDurationS":228093.81403760775,"totalInputTokens":98237249.5,"totalOutputTokens":1413585,"total_cost_usd":8.86,"total_input_tokens":98237249.5,"total_output_tokens":1413585,"total_runtime_s":228093.81403760775,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.547984743567735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-bf8acf5fcd2d383e5d5f","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d2755e31fd8a24b80a78a03f","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1492606367247545,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":950,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[474];split=unlicensed;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":950,"metadata":{"avgScore":0.1492606367247545,"avg_score":0.1492606367247545,"avg_score_percent":14.92606367247545,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":25.31,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":274979228.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":3748463.1666666665,"passRate":0.06382978723404255,"pass_rate":0.06382978723404255,"pass_rate_percent":6.382978723404255,"passes":14,"runs":216,"runtime":442594.7348369967,"source_flags":[],"source_harness":"openclaw","source_row_index":474,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":25.31,"totalDurationS":442594.7348369967,"totalInputTokens":274979228.6666667,"totalOutputTokens":3748463.1666666665,"total_cost_usd":25.31,"total_input_tokens":274979228.6666667,"total_output_tokens":3748463.1666666665,"total_runtime_s":442594.7348369967,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.92606367247545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-10055a13eedb4d8abe7a","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a347974681c81045fd90345e","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.012792105263157895,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":350,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[174];split=full/last-exam;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":350,"metadata":{"avgScore":0.012792105263157895,"avg_score":0.012792105263157895,"avg_score_percent":1.2792105263157894,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":9.15,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":87693128.5,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":1462761.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":47,"runtime":159244.5490859223,"source_flags":[],"source_harness":"openclaw","source_row_index":174,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":9.15,"totalDurationS":159244.5490859223,"totalInputTokens":87693128.5,"totalOutputTokens":1462761.8333333333,"total_cost_usd":9.15,"total_input_tokens":87693128.5,"total_output_tokens":1462761.8333333333,"total_runtime_s":159244.5490859223,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":1.2792105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-77c7bb4117a8c878663d","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7fea35bebb3ed684d04b5925","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1464903265101336,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":834,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[416];split=linux_only;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":834,"metadata":{"avgScore":0.1464903265101336,"avg_score":0.1464903265101336,"avg_score_percent":14.64903265101336,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.15,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":242674106.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":3192715.3333333335,"passRate":0.05714285714285714,"pass_rate":0.05714285714285714,"pass_rate_percent":5.714285714285714,"passes":9,"runs":167,"runtime":190567.1761067583,"source_flags":[],"source_harness":"openclaw","source_row_index":416,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":22.15,"totalDurationS":190567.1761067583,"totalInputTokens":242674106.16666666,"totalOutputTokens":3192715.3333333335,"total_cost_usd":22.15,"total_input_tokens":242674106.16666666,"total_output_tokens":3192715.3333333335,"total_runtime_s":190567.1761067583,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.64903265101336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0e3922f41ad64ef9e47d","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6f5dc2dd04061f25bc5b0129","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.14187993275125252,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":586,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[292];split=full/overall;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":586,"metadata":{"avgScore":0.14187993275125252,"avg_score":0.14187993275125252,"avg_score_percent":14.187993275125251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.74,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":277209591.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":4379146.166666667,"passRate":0.05921052631578947,"pass_rate":0.05921052631578947,"pass_rate_percent":5.921052631578947,"passes":14,"runs":223,"runtime":520646.2217991107,"source_flags":[],"source_harness":"openclaw","source_row_index":292,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":26.74,"totalDurationS":520646.2217991107,"totalInputTokens":277209591.6666667,"totalOutputTokens":4379146.166666667,"total_cost_usd":26.74,"total_input_tokens":277209591.6666667,"total_output_tokens":4379146.166666667,"total_runtime_s":520646.2217991107,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.187993275125251,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d974e6565a7ce42668e3","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6e98581ea7490b6ac27c6613","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08225000000000002,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1070,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[534];split=unlicensed/full-spectrum;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1070,"metadata":{"avgScore":0.08225000000000002,"avg_score":0.08225000000000002,"avg_score_percent":8.225000000000001,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.45,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":92955565.16666667,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":1210495.1666666667,"passRate":0.04,"pass_rate":0.04,"pass_rate_percent":4.0,"passes":2,"runs":79,"runtime":127004.4180870185,"source_flags":[],"source_harness":"openclaw","source_row_index":534,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":8.45,"totalDurationS":127004.4180870185,"totalInputTokens":92955565.16666667,"totalOutputTokens":1210495.1666666667,"total_cost_usd":8.45,"total_input_tokens":92955565.16666667,"total_output_tokens":1210495.1666666667,"total_runtime_s":127004.4180870185,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.225000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-acc7483a058d79ac1b21","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6188adf374b274b0db659d8e","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08422727272727273,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":234,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[116];split=full/full-spectrum;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":234,"metadata":{"avgScore":0.08422727272727273,"avg_score":0.08422727272727273,"avg_score_percent":8.422727272727274,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":9.56,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":95157073.66666667,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":1732033.6666666667,"passRate":0.03636363636363636,"pass_rate":0.03636363636363636,"pass_rate_percent":3.6363636363636362,"passes":2,"runs":85,"runtime":166639.72889767349,"source_flags":[],"source_harness":"openclaw","source_row_index":116,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":9.56,"totalDurationS":166639.72889767349,"totalInputTokens":95157073.66666667,"totalOutputTokens":1732033.6666666667,"total_cost_usd":9.56,"total_input_tokens":95157073.66666667,"total_output_tokens":1732033.6666666667,"total_runtime_s":166639.72889767349,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.422727272727274,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ce47eb639a6d3012a744","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-51cc0ce91c449bb46a93f031","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24547984743567736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":466,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[232];split=full/near-term;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":466,"metadata":{"avgScore":0.24547984743567736,"avg_score":0.24547984743567736,"avg_score_percent":24.547984743567735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.86,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98237249.5,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":1413585,"passRate":0.1044776119402985,"pass_rate":0.1044776119402985,"pass_rate_percent":10.44776119402985,"passes":12,"runs":105,"runtime":228093.81403760775,"source_flags":[],"source_harness":"openclaw","source_row_index":232,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":8.86,"totalDurationS":228093.81403760775,"totalInputTokens":98237249.5,"totalOutputTokens":1413585,"total_cost_usd":8.86,"total_input_tokens":98237249.5,"total_output_tokens":1413585,"total_runtime_s":228093.81403760775,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.547984743567735,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f635e21b972f18141668","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4ae705a0163bb00669aaadae","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.04727272727272727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":694,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[346];split=licensed;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":694,"metadata":{"avgScore":0.04727272727272727,"avg_score":0.04727272727272727,"avg_score_percent":4.7272727272727275,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1.43,"costSource":"list","cost_source":"list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2230363,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":630683,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":78051.486962114,"source_flags":[],"source_harness":"openclaw","source_row_index":346,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":1.43,"totalDurationS":78051.486962114,"totalInputTokens":2230363,"totalOutputTokens":630683,"total_cost_usd":1.43,"total_input_tokens":2230363,"total_output_tokens":630683,"total_runtime_s":78051.486962114,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.7272727272727275,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7a64129747aac7294e4c","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3f0398941161bc6aa6b6e9d0","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1492606367247545,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1422,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[710];split=unlicensed/overall;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1422,"metadata":{"avgScore":0.1492606367247545,"avg_score":0.1492606367247545,"avg_score_percent":14.92606367247545,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":25.31,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":274979228.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":3748463.1666666665,"passRate":0.06382978723404255,"pass_rate":0.06382978723404255,"pass_rate_percent":6.382978723404255,"passes":14,"runs":216,"runtime":442594.7348369967,"source_flags":[],"source_harness":"openclaw","source_row_index":710,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":25.31,"totalDurationS":442594.7348369967,"totalInputTokens":274979228.6666667,"totalOutputTokens":3748463.1666666665,"total_cost_usd":25.31,"total_input_tokens":274979228.6666667,"total_output_tokens":3748463.1666666665,"total_runtime_s":442594.7348369967,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.92606367247545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-301fd3b0d294351ad1f9","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-34bb7cd010a3aba12f772841","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.14187993275125252,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070c3da4f9b7a530d3e39eefe407cacd1da203b7","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":114,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[56];split=full;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":114,"metadata":{"avgScore":0.14187993275125252,"avg_score":0.14187993275125252,"avg_score_percent":14.187993275125251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.74,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":277209591.6666667,"metric_id":"avg_score","metric_role":"secondary","model":"minimax-m2-7","output_tokens":4379146.166666667,"passRate":0.05921052631578947,"pass_rate":0.05921052631578947,"pass_rate_percent":5.921052631578947,"passes":14,"runs":223,"runtime":520646.2217991107,"source_flags":[],"source_harness":"openclaw","source_row_index":56,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":26.74,"totalDurationS":520646.2217991107,"totalInputTokens":277209591.6666667,"totalOutputTokens":4379146.166666667,"total_cost_usd":26.74,"total_input_tokens":277209591.6666667,"total_output_tokens":4379146.166666667,"total_runtime_s":520646.2217991107,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":14.187993275125251,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3842d9e7e0a196c25b49","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-dce7b6bbc86087f8fe731608","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1185,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[592];split=unlicensed/last-exam;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1185,"metadata":{"avgScore":0.016203333333333333,"avg_score":0.016203333333333333,"avg_score_percent":1.6203333333333334,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.37,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":86177572.5,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":1194918.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":43,"runtime":109949.42864178025,"source_flags":[],"source_harness":"openclaw","source_row_index":592,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":8.37,"totalDurationS":109949.42864178025,"totalInputTokens":86177572.5,"totalOutputTokens":1194918.8333333333,"total_cost_usd":8.37,"total_input_tokens":86177572.5,"total_output_tokens":1194918.8333333333,"total_runtime_s":109949.42864178025,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c1018ef32568dd3d7f2","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b26b43ed50ea06b3429cc82f","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.05921052631578947,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":113,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[56];split=full;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":113,"metadata":{"avgScore":0.14187993275125252,"avg_score":0.14187993275125252,"avg_score_percent":14.187993275125251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.74,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":277209591.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":4379146.166666667,"passRate":0.05921052631578947,"pass_rate":0.05921052631578947,"pass_rate_percent":5.921052631578947,"passes":14,"runs":223,"runtime":520646.2217991107,"source_flags":[],"source_harness":"openclaw","source_row_index":56,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":26.74,"totalDurationS":520646.2217991107,"totalInputTokens":277209591.6666667,"totalOutputTokens":4379146.166666667,"total_cost_usd":26.74,"total_input_tokens":277209591.6666667,"total_output_tokens":4379146.166666667,"total_runtime_s":520646.2217991107,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":5.921052631578947,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4802929bc5b012e6e52d","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-afcab69be4539d6de9a9d23a","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1044776119402985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":465,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[232];split=full/near-term;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":465,"metadata":{"avgScore":0.24547984743567736,"avg_score":0.24547984743567736,"avg_score_percent":24.547984743567735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.86,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98237249.5,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":1413585,"passRate":0.1044776119402985,"pass_rate":0.1044776119402985,"pass_rate_percent":10.44776119402985,"passes":12,"runs":105,"runtime":228093.81403760775,"source_flags":[],"source_harness":"openclaw","source_row_index":232,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":8.86,"totalDurationS":228093.81403760775,"totalInputTokens":98237249.5,"totalOutputTokens":1413585,"total_cost_usd":8.86,"total_input_tokens":98237249.5,"total_output_tokens":1413585,"total_runtime_s":228093.81403760775,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.44776119402985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-dec4fc72173262a65517","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aa46d4e0866a6ec53b2673f1","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.05714285714285714,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":833,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[416];split=linux_only;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":833,"metadata":{"avgScore":0.1464903265101336,"avg_score":0.1464903265101336,"avg_score_percent":14.64903265101336,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":22.15,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":242674106.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":3192715.3333333335,"passRate":0.05714285714285714,"pass_rate":0.05714285714285714,"pass_rate_percent":5.714285714285714,"passes":9,"runs":167,"runtime":190567.1761067583,"source_flags":[],"source_harness":"openclaw","source_row_index":416,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":22.15,"totalDurationS":190567.1761067583,"totalInputTokens":242674106.16666666,"totalOutputTokens":3192715.3333333335,"total_cost_usd":22.15,"total_input_tokens":242674106.16666666,"total_output_tokens":3192715.3333333335,"total_runtime_s":190567.1761067583,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":5.714285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-954205d9ae004e6a88e9","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9e6388595549f1487f6184ed","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":693,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[346];split=licensed;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":693,"metadata":{"avgScore":0.04727272727272727,"avg_score":0.04727272727272727,"avg_score_percent":4.7272727272727275,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":1.43,"costSource":"list","cost_source":"list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":2230363,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":630683,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":78051.486962114,"source_flags":[],"source_harness":"openclaw","source_row_index":346,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":1.43,"totalDurationS":78051.486962114,"totalInputTokens":2230363,"totalOutputTokens":630683,"total_cost_usd":1.43,"total_input_tokens":2230363,"total_output_tokens":630683,"total_runtime_s":78051.486962114,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f7c1de93f20b5ff0e312","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-913cde44d362651a3ba38d8e","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.04,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1069,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[534];split=unlicensed/full-spectrum;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1069,"metadata":{"avgScore":0.08225000000000002,"avg_score":0.08225000000000002,"avg_score_percent":8.225000000000001,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.45,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":92955565.16666667,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":1210495.1666666667,"passRate":0.04,"pass_rate":0.04,"pass_rate_percent":4.0,"passes":2,"runs":79,"runtime":127004.4180870185,"source_flags":[],"source_harness":"openclaw","source_row_index":534,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":8.45,"totalDurationS":127004.4180870185,"totalInputTokens":92955565.16666667,"totalOutputTokens":1210495.1666666667,"total_cost_usd":8.45,"total_input_tokens":92955565.16666667,"total_output_tokens":1210495.1666666667,"total_runtime_s":127004.4180870185,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4cec78123d33134a01a1","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8ad91645b23ea4c61e7e1064","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":349,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[174];split=full/last-exam;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":349,"metadata":{"avgScore":0.012792105263157895,"avg_score":0.012792105263157895,"avg_score_percent":1.2792105263157894,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":9.15,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":87693128.5,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":1462761.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":47,"runtime":159244.5490859223,"source_flags":[],"source_harness":"openclaw","source_row_index":174,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":9.15,"totalDurationS":159244.5490859223,"totalInputTokens":87693128.5,"totalOutputTokens":1462761.8333333333,"total_cost_usd":9.15,"total_input_tokens":87693128.5,"total_output_tokens":1462761.8333333333,"total_runtime_s":159244.5490859223,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-08fb091e0a2d7c2358f1","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-86009f935aa8d4076be80a18","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06382978723404255,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":949,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[474];split=unlicensed;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":949,"metadata":{"avgScore":0.1492606367247545,"avg_score":0.1492606367247545,"avg_score_percent":14.92606367247545,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":25.31,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":274979228.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":3748463.1666666665,"passRate":0.06382978723404255,"pass_rate":0.06382978723404255,"pass_rate_percent":6.382978723404255,"passes":14,"runs":216,"runtime":442594.7348369967,"source_flags":[],"source_harness":"openclaw","source_row_index":474,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":25.31,"totalDurationS":442594.7348369967,"totalInputTokens":274979228.6666667,"totalOutputTokens":3748463.1666666665,"total_cost_usd":25.31,"total_input_tokens":274979228.6666667,"total_output_tokens":3748463.1666666665,"total_runtime_s":442594.7348369967,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.382978723404255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3e1a269d604d1cea102f","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7df899d18ee17c6b998b3ef2","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.03636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":233,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[116];split=full/full-spectrum;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":233,"metadata":{"avgScore":0.08422727272727273,"avg_score":0.08422727272727273,"avg_score_percent":8.422727272727274,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":9.56,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":95157073.66666667,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":1732033.6666666667,"passRate":0.03636363636363636,"pass_rate":0.03636363636363636,"pass_rate_percent":3.6363636363636362,"passes":2,"runs":85,"runtime":166639.72889767349,"source_flags":[],"source_harness":"openclaw","source_row_index":116,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":9.56,"totalDurationS":166639.72889767349,"totalInputTokens":95157073.66666667,"totalOutputTokens":1732033.6666666667,"total_cost_usd":9.56,"total_input_tokens":95157073.66666667,"total_output_tokens":1732033.6666666667,"total_runtime_s":166639.72889767349,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":3.6363636363636362,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4e44dea33a50ea3ba649","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7985e5a89b9af2cc6c8f0336","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.05921052631578947,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":585,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[292];split=full/overall;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":585,"metadata":{"avgScore":0.14187993275125252,"avg_score":0.14187993275125252,"avg_score_percent":14.187993275125251,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":26.74,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":277209591.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":4379146.166666667,"passRate":0.05921052631578947,"pass_rate":0.05921052631578947,"pass_rate_percent":5.921052631578947,"passes":14,"runs":223,"runtime":520646.2217991107,"source_flags":[],"source_harness":"openclaw","source_row_index":292,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":26.74,"totalDurationS":520646.2217991107,"totalInputTokens":277209591.6666667,"totalOutputTokens":4379146.166666667,"total_cost_usd":26.74,"total_input_tokens":277209591.6666667,"total_output_tokens":4379146.166666667,"total_runtime_s":520646.2217991107,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":5.921052631578947,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-260785e2f48d0d0f2fa8","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5fc65d84813a2ff1b0a93134","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06382978723404255,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1421,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[710];split=unlicensed/overall;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1421,"metadata":{"avgScore":0.1492606367247545,"avg_score":0.1492606367247545,"avg_score_percent":14.92606367247545,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":25.31,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":274979228.6666667,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":3748463.1666666665,"passRate":0.06382978723404255,"pass_rate":0.06382978723404255,"pass_rate_percent":6.382978723404255,"passes":14,"runs":216,"runtime":442594.7348369967,"source_flags":[],"source_harness":"openclaw","source_row_index":710,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":25.31,"totalDurationS":442594.7348369967,"totalInputTokens":274979228.6666667,"totalOutputTokens":3748463.1666666665,"total_cost_usd":25.31,"total_input_tokens":274979228.6666667,"total_output_tokens":3748463.1666666665,"total_runtime_s":442594.7348369967,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.382978723404255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-90c7fac5f62d9425fd9e","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2d74ff1d327f419326f488d4","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"MiniMax M2.7","modelRef":"minimax-m2-7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1044776119402985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b4a7fd0ccfd9adedff9547983ca275e38251f2a","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1301,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[650];split=unlicensed/near-term;harness=openclaw;model=minimax-m2-7;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1301,"metadata":{"avgScore":0.24547984743567736,"avg_score":0.24547984743567736,"avg_score_percent":24.547984743567735,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":8.86,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":98237249.5,"metric_id":"pass_rate","metric_role":"primary","model":"minimax-m2-7","output_tokens":1413585,"passRate":0.1044776119402985,"pass_rate":0.1044776119402985,"pass_rate_percent":10.44776119402985,"passes":12,"runs":105,"runtime":228093.81403760775,"source_flags":[],"source_harness":"openclaw","source_row_index":650,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":8.86,"totalDurationS":228093.81403760775,"totalInputTokens":98237249.5,"totalOutputTokens":1413585,"total_cost_usd":8.86,"total_input_tokens":98237249.5,"total_output_tokens":1413585,"total_runtime_s":228093.81403760775,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.44776119402985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d67d25ed20c0c27a9f58","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f428766438793db87c5c47c2","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4354068132927283,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":764,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[381];split=linux_only;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":764,"metadata":{"avgScore":0.4354068132927283,"avg_score":0.4354068132927283,"avg_score_percent":43.54068132927283,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":1627349109,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":10241576,"passRate":0.19047619047619047,"pass_rate":0.19047619047619047,"pass_rate_percent":19.047619047619047,"passes":20,"runs":104,"runtime":583968.235,"source_flags":[],"source_harness":"claude_code","source_row_index":381,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":null,"totalDurationS":583968.235,"totalInputTokens":1627349109,"totalOutputTokens":10241576,"total_cost_usd":null,"total_input_tokens":1627349109,"total_output_tokens":10241576,"total_runtime_s":583968.235,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.54068132927283,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6c7d38c10c8772e42e3d","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-e8cf25002efedec4450a5c8b","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.14726424569791585,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":264,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[131];split=full/last-exam;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":264,"metadata":{"avgScore":0.14726424569791585,"avg_score":0.14726424569791585,"avg_score_percent":14.726424569791586,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":894801044,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":4063160,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":35,"runtime":410964.91899999994,"source_flags":[],"source_harness":"claude_code","source_row_index":131,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":35,"tokens":null,"totalCostUsd":null,"totalDurationS":410964.91899999994,"totalInputTokens":894801044,"totalOutputTokens":4063160,"total_cost_usd":null,"total_input_tokens":894801044,"total_output_tokens":4063160,"total_runtime_s":410964.91899999994,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":14.726424569791586,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d2817bdb2d75b91bad2","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-da1a3150214aa5ebded8ccc0","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.13893693343958233,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1104,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[551];split=unlicensed/last-exam;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1104,"metadata":{"avgScore":0.13893693343958233,"avg_score":0.13893693343958233,"avg_score_percent":13.893693343958233,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":808120996,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":3483338,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":28,"runtime":324532.41500000004,"source_flags":[],"source_harness":"claude_code","source_row_index":551,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":null,"totalDurationS":324532.41500000004,"totalInputTokens":808120996,"totalOutputTokens":3483338,"total_cost_usd":null,"total_input_tokens":808120996,"total_output_tokens":3483338,"total_runtime_s":324532.41500000004,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.893693343958233,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2c03ba3b24dae470bd0d","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d2f2016a94549d0ba870c874","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.6002555857554045,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1262,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[630];split=unlicensed/near-term;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1262,"metadata":{"avgScore":0.6002555857554045,"avg_score":0.6002555857554045,"avg_score_percent":60.02555857554045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":684638103,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":6102227,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":298438.2150000001,"source_flags":[],"source_harness":"claude_code","source_row_index":630,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":null,"totalDurationS":298438.2150000001,"totalInputTokens":684638103,"totalOutputTokens":6102227,"total_cost_usd":null,"total_input_tokens":684638103,"total_output_tokens":6102227,"total_runtime_s":298438.2150000001,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.02555857554045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4f9e86b28cb948db4394","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b81e78692fde370d4e06601d","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4459353752236476,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":882,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[440];split=unlicensed;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":882,"metadata":{"avgScore":0.4459353752236476,"avg_score":0.4459353752236476,"avg_score_percent":44.59353752236476,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2489228093,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":14174887,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":138,"runtime":873344.9099999998,"source_flags":[],"source_harness":"claude_code","source_row_index":440,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":null,"totalDurationS":873344.9099999998,"totalInputTokens":2489228093,"totalOutputTokens":14174887,"total_cost_usd":null,"total_input_tokens":2489228093,"total_output_tokens":14174887,"total_runtime_s":873344.9099999998,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.59353752236476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8589904b3cc9a25aa683","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b488cf92e288ee999560d6f3","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4459353752236476,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1354,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[676];split=unlicensed/overall;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1354,"metadata":{"avgScore":0.4459353752236476,"avg_score":0.4459353752236476,"avg_score_percent":44.59353752236476,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2489228093,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":14174887,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":138,"runtime":873344.9099999998,"source_flags":[],"source_harness":"claude_code","source_row_index":676,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":null,"totalDurationS":873344.9099999998,"totalInputTokens":2489228093,"totalOutputTokens":14174887,"total_cost_usd":null,"total_input_tokens":2489228093,"total_output_tokens":14174887,"total_runtime_s":873344.9099999998,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.59353752236476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5793b9b2bb9a478f7f6f","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-adc96de822791ee08490bce9","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.428847508157024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":516,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[257];split=full/overall;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":516,"metadata":{"avgScore":0.428847508157024,"avg_score":0.428847508157024,"avg_score_percent":42.8847508157024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2742661908,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":14977273,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":148,"runtime":1000718.9020000002,"source_flags":[],"source_harness":"claude_code","source_row_index":257,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":null,"totalDurationS":1000718.9020000002,"totalInputTokens":2742661908,"totalOutputTokens":14977273,"total_cost_usd":null,"total_input_tokens":2742661908,"total_output_tokens":14977273,"total_runtime_s":1000718.9020000002,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.8847508157024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a455f383cff2bbb1c9d8","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-66eaecda79f6df347ddf4b3d","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.428847508157024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":44,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[21];split=full;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":44,"metadata":{"avgScore":0.428847508157024,"avg_score":0.428847508157024,"avg_score_percent":42.8847508157024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2742661908,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":14977273,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":148,"runtime":1000718.9020000002,"source_flags":[],"source_harness":"claude_code","source_row_index":21,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":null,"totalDurationS":1000718.9020000002,"totalInputTokens":2742661908,"totalOutputTokens":14977273,"total_cost_usd":null,"total_input_tokens":2742661908,"total_output_tokens":14977273,"total_runtime_s":1000718.9020000002,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.8847508157024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cbd5f567103aa344b386","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-541a42c369df819bcfcc2767","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.43039726926162564,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":976,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[487];split=unlicensed/full-spectrum;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":976,"metadata":{"avgScore":0.43039726926162564,"avg_score":0.43039726926162564,"avg_score_percent":43.039726926162565,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":1068561142,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":5121884,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":50,"runtime":290651.849,"source_flags":[],"source_harness":"claude_code","source_row_index":487,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":null,"totalDurationS":290651.849,"totalInputTokens":1068561142,"totalOutputTokens":5121884,"total_cost_usd":null,"total_input_tokens":1068561142,"total_output_tokens":5121884,"total_runtime_s":290651.849,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.039726926162565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-44152a7f1d295d551156","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-3c4b438fb13b1dc33fc88fa5","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.6002555857554045,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":426,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[212];split=full/near-term;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":426,"metadata":{"avgScore":0.6002555857554045,"avg_score":0.6002555857554045,"avg_score_percent":60.02555857554045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":684638103,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":6102227,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":298438.2150000001,"source_flags":[],"source_harness":"claude_code","source_row_index":212,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":null,"totalDurationS":298438.2150000001,"totalInputTokens":684638103,"totalOutputTokens":6102227,"total_cost_usd":null,"total_input_tokens":684638103,"total_output_tokens":6102227,"total_runtime_s":298438.2150000001,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.02555857554045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2506c6454f79f51c0809","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-22da0551701e0ac0cfcd48d3","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.4254520629651143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":142,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[70];split=full/full-spectrum;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":142,"metadata":{"avgScore":0.4254520629651143,"avg_score":0.4254520629651143,"avg_score_percent":42.54520629651143,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":1246333555,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":5369786,"passRate":0.21818181818181817,"pass_rate":0.21818181818181817,"pass_rate_percent":21.818181818181817,"passes":12,"runs":55,"runtime":356231.91900000005,"source_flags":[],"source_harness":"claude_code","source_row_index":70,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":null,"totalDurationS":356231.91900000005,"totalInputTokens":1246333555,"totalOutputTokens":5369786,"total_cost_usd":null,"total_input_tokens":1246333555,"total_output_tokens":5369786,"total_runtime_s":356231.91900000005,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":42.54520629651143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9626f227e1d6a42525aa","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-08d1a3c1d28245166029a134","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.20981212121212123,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c080f9a19a77e2f847ba8f10900d15aed510935","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":626,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[312];split=licensed;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":626,"metadata":{"avgScore":0.20981212121212123,"avg_score":0.20981212121212123,"avg_score_percent":20.981212121212124,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":253433815,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-27b","output_tokens":802386,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":127373.992,"source_flags":[],"source_harness":"claude_code","source_row_index":312,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":null,"totalDurationS":127373.992,"totalInputTokens":253433815,"totalOutputTokens":802386,"total_cost_usd":null,"total_input_tokens":253433815,"total_output_tokens":802386,"total_runtime_s":127373.992,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.981212121212124,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4b6ea50fb2a96f0b4829","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-eb98d81d07aa8f8957bc44ed","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.31343283582089554,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1261,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[630];split=unlicensed/near-term;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1261,"metadata":{"avgScore":0.6002555857554045,"avg_score":0.6002555857554045,"avg_score_percent":60.02555857554045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":684638103,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":6102227,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":298438.2150000001,"source_flags":[],"source_harness":"claude_code","source_row_index":630,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":null,"totalDurationS":298438.2150000001,"totalInputTokens":684638103,"totalOutputTokens":6102227,"total_cost_usd":null,"total_input_tokens":684638103,"total_output_tokens":6102227,"total_runtime_s":298438.2150000001,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.343283582089555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-558395a82b1ef4b9e9e1","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-db99a36538cf40dcb85e252e","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":625,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[312];split=licensed;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":625,"metadata":{"avgScore":0.20981212121212123,"avg_score":0.20981212121212123,"avg_score_percent":20.981212121212124,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":253433815,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":802386,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":127373.992,"source_flags":[],"source_harness":"claude_code","source_row_index":312,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":null,"totalDurationS":127373.992,"totalInputTokens":253433815,"totalOutputTokens":802386,"total_cost_usd":null,"total_input_tokens":253433815,"total_output_tokens":802386,"total_runtime_s":127373.992,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d714797a13251103e664","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d27be26eedf3b5cf1a8bdce3","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.22,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":975,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[487];split=unlicensed/full-spectrum;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":975,"metadata":{"avgScore":0.43039726926162564,"avg_score":0.43039726926162564,"avg_score_percent":43.039726926162565,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":1068561142,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":5121884,"passRate":0.22,"pass_rate":0.22,"pass_rate_percent":22.0,"passes":11,"runs":50,"runtime":290651.849,"source_flags":[],"source_harness":"claude_code","source_row_index":487,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":50,"tokens":null,"totalCostUsd":null,"totalDurationS":290651.849,"totalInputTokens":1068561142,"totalOutputTokens":5121884,"total_cost_usd":null,"total_input_tokens":1068561142,"total_output_tokens":5121884,"total_runtime_s":290651.849,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-2e99086eb3328fa411cc","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-cd3bd68a3a538ed1a5ee8599","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.03333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1103,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[551];split=unlicensed/last-exam;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1103,"metadata":{"avgScore":0.13893693343958233,"avg_score":0.13893693343958233,"avg_score_percent":13.893693343958233,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":808120996,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":3483338,"passRate":0.03333333333333333,"pass_rate":0.03333333333333333,"pass_rate_percent":3.3333333333333335,"passes":1,"runs":28,"runtime":324532.41500000004,"source_flags":[],"source_harness":"claude_code","source_row_index":551,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":28,"tokens":null,"totalCostUsd":null,"totalDurationS":324532.41500000004,"totalInputTokens":808120996,"totalOutputTokens":3483338,"total_cost_usd":null,"total_input_tokens":808120996,"total_output_tokens":3483338,"total_runtime_s":324532.41500000004,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":3.3333333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-967aa1792e656636ee7e","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-adbf21ba10ee7b24cceee447","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.31343283582089554,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":425,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[212];split=full/near-term;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":425,"metadata":{"avgScore":0.6002555857554045,"avg_score":0.6002555857554045,"avg_score_percent":60.02555857554045,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":684638103,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":6102227,"passRate":0.31343283582089554,"pass_rate":0.31343283582089554,"pass_rate_percent":31.343283582089555,"passes":21,"runs":66,"runtime":298438.2150000001,"source_flags":[],"source_harness":"claude_code","source_row_index":212,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":66,"tokens":null,"totalCostUsd":null,"totalDurationS":298438.2150000001,"totalInputTokens":684638103,"totalOutputTokens":6102227,"total_cost_usd":null,"total_input_tokens":684638103,"total_output_tokens":6102227,"total_runtime_s":298438.2150000001,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":31.343283582089555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d48bc0a47d06be7e8292","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-aab6f38697649d96f98c4768","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2198581560283688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1353,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[676];split=unlicensed/overall;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1353,"metadata":{"avgScore":0.4459353752236476,"avg_score":0.4459353752236476,"avg_score_percent":44.59353752236476,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2489228093,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":14174887,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":138,"runtime":873344.9099999998,"source_flags":[],"source_harness":"claude_code","source_row_index":676,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":null,"totalDurationS":873344.9099999998,"totalInputTokens":2489228093,"totalOutputTokens":14174887,"total_cost_usd":null,"total_input_tokens":2489228093,"total_output_tokens":14174887,"total_runtime_s":873344.9099999998,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.98581560283688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fe215aac4a38730f5128","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-9803ba36ebe45f006d7fc6d7","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.21818181818181817,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":141,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[70];split=full/full-spectrum;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":141,"metadata":{"avgScore":0.4254520629651143,"avg_score":0.4254520629651143,"avg_score_percent":42.54520629651143,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":1246333555,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":5369786,"passRate":0.21818181818181817,"pass_rate":0.21818181818181817,"pass_rate_percent":21.818181818181817,"passes":12,"runs":55,"runtime":356231.91900000005,"source_flags":[],"source_harness":"claude_code","source_row_index":70,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":55,"tokens":null,"totalCostUsd":null,"totalDurationS":356231.91900000005,"totalInputTokens":1246333555,"totalOutputTokens":5369786,"total_cost_usd":null,"total_input_tokens":1246333555,"total_output_tokens":5369786,"total_runtime_s":356231.91900000005,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.818181818181817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7045121a82073fe6d909","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7bfe1263c5850ff90effdddc","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.02631578947368421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":263,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[131];split=full/last-exam;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":263,"metadata":{"avgScore":0.14726424569791585,"avg_score":0.14726424569791585,"avg_score_percent":14.726424569791586,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":894801044,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":4063160,"passRate":0.02631578947368421,"pass_rate":0.02631578947368421,"pass_rate_percent":2.631578947368421,"passes":1,"runs":35,"runtime":410964.91899999994,"source_flags":[],"source_harness":"claude_code","source_row_index":131,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":35,"tokens":null,"totalCostUsd":null,"totalDurationS":410964.91899999994,"totalInputTokens":894801044,"totalOutputTokens":4063160,"total_cost_usd":null,"total_input_tokens":894801044,"total_output_tokens":4063160,"total_runtime_s":410964.91899999994,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":2.631578947368421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c395295a118799ba4912","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-603b44dc756f66a2588b7f35","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.19047619047619047,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":763,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[381];split=linux_only;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":763,"metadata":{"avgScore":0.4354068132927283,"avg_score":0.4354068132927283,"avg_score_percent":43.54068132927283,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":1627349109,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":10241576,"passRate":0.19047619047619047,"pass_rate":0.19047619047619047,"pass_rate_percent":19.047619047619047,"passes":20,"runs":104,"runtime":583968.235,"source_flags":[],"source_harness":"claude_code","source_row_index":381,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":null,"totalDurationS":583968.235,"totalInputTokens":1627349109,"totalOutputTokens":10241576,"total_cost_usd":null,"total_input_tokens":1627349109,"total_output_tokens":10241576,"total_runtime_s":583968.235,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":19.047619047619047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-edca4b399c29f7897149","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-55b6225a7777ddab96c3b232","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.20394736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":43,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[21];split=full;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":43,"metadata":{"avgScore":0.428847508157024,"avg_score":0.428847508157024,"avg_score_percent":42.8847508157024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2742661908,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":14977273,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":148,"runtime":1000718.9020000002,"source_flags":[],"source_harness":"claude_code","source_row_index":21,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":null,"totalDurationS":1000718.9020000002,"totalInputTokens":2742661908,"totalOutputTokens":14977273,"total_cost_usd":null,"total_input_tokens":2742661908,"total_output_tokens":14977273,"total_runtime_s":1000718.9020000002,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.394736842105264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-02e3f5d966bb9d8d819e","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-13e19d2a80f85127cc568923","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.20394736842105263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":515,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[257];split=full/overall;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":515,"metadata":{"avgScore":0.428847508157024,"avg_score":0.428847508157024,"avg_score_percent":42.8847508157024,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2742661908,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":14977273,"passRate":0.20394736842105263,"pass_rate":0.20394736842105263,"pass_rate_percent":20.394736842105264,"passes":31,"runs":148,"runtime":1000718.9020000002,"source_flags":[],"source_harness":"claude_code","source_row_index":257,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":null,"totalDurationS":1000718.9020000002,"totalInputTokens":2742661908,"totalOutputTokens":14977273,"total_cost_usd":null,"total_input_tokens":2742661908,"total_output_tokens":14977273,"total_runtime_s":1000718.9020000002,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":20.394736842105264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-fc062bee17b58f3361a7","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0b0868699b36975af33e4514","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 27B","modelRef":"qwen-3-8-27b","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":null,"evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"reasoning-max","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2198581560283688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7ca9d55393a076f0e4f9d71e703e356e28123cd","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":881,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[440];split=unlicensed;harness=claude_code;model=qwen-3-8-27b;variant=reasoning-max;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":881,"metadata":{"avgScore":0.4459353752236476,"avg_score":0.4459353752236476,"avg_score_percent":44.59353752236476,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":null,"costSource":null,"cost_source":null,"harness":"claude_code","harness_id":"claude-code","harness_variant":"reasoning-max","input_tokens":2489228093,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-27b","output_tokens":14174887,"passRate":0.2198581560283688,"pass_rate":0.2198581560283688,"pass_rate_percent":21.98581560283688,"passes":31,"runs":138,"runtime":873344.9099999998,"source_flags":[],"source_harness":"claude_code","source_row_index":440,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":null,"totalDurationS":873344.9099999998,"totalInputTokens":2489228093,"totalOutputTokens":14174887,"total_cost_usd":null,"total_input_tokens":2489228093,"total_output_tokens":14174887,"total_runtime_s":873344.9099999998,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":21.98581560283688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-648184fb13c78593cbf1","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-f8e4c4942c68ebeb56ac07c0","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5526026381504506,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1312,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[655];split=unlicensed/overall;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1312,"metadata":{"avgScore":0.5526026381504506,"avg_score":0.5526026381504506,"avg_score_percent":55.260263815045064,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":406.16,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":977228809,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":10754816,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":138,"runtime":622739.765,"source_flags":[],"source_harness":"claude_code","source_row_index":655,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":406.16,"totalDurationS":622739.765,"totalInputTokens":977228809,"totalOutputTokens":10754816,"total_cost_usd":406.16,"total_input_tokens":977228809,"total_output_tokens":10754816,"total_runtime_s":622739.765,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":55.260263815045064,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4b9a1de9e5ed75817f79","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-d2ddbe8f0f6820cdad76f66a","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.46354442255769973,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":124,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[61];split=full/full-spectrum;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":124,"metadata":{"avgScore":0.46354442255769973,"avg_score":0.46354442255769973,"avg_score_percent":46.354442255769975,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":149.3,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":392062495,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":3623554,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":53,"runtime":207159.591,"source_flags":[],"source_harness":"claude_code","source_row_index":61,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":149.3,"totalDurationS":207159.591,"totalInputTokens":392062495,"totalOutputTokens":3623554,"total_cost_usd":149.3,"total_input_tokens":392062495,"total_output_tokens":3623554,"total_runtime_s":207159.591,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.354442255769975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f1f5bdb904004d6ea293","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-aba30c1dc747f8011febe368","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5248561340297907,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":478,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[238];split=full/overall;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":478,"metadata":{"avgScore":0.5248561340297907,"avg_score":0.5248561340297907,"avg_score_percent":52.48561340297907,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":486.47,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1234201310,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":12430863,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":148,"runtime":729633.2999999999,"source_flags":[],"source_harness":"claude_code","source_row_index":238,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":486.47,"totalDurationS":729633.2999999999,"totalInputTokens":1234201310,"totalOutputTokens":12430863,"total_cost_usd":486.47,"total_input_tokens":1234201310,"total_output_tokens":12430863,"total_runtime_s":729633.2999999999,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.48561340297907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d13b08ff9a4578934733","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-917614acbc3f502e032a5b0d","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.46727232361384374,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":958,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[478];split=unlicensed/full-spectrum;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":958,"metadata":{"avgScore":0.46727232361384374,"avg_score":0.46727232361384374,"avg_score_percent":46.727232361384374,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":120.6,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":298336188,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":3103340,"passRate":0.24,"pass_rate":0.24,"pass_rate_percent":24.0,"passes":12,"runs":48,"runtime":161452.09199999998,"source_flags":[],"source_harness":"claude_code","source_row_index":478,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":120.6,"totalDurationS":161452.09199999998,"totalInputTokens":298336188,"totalOutputTokens":3103340,"total_cost_usd":120.6,"total_input_tokens":298336188,"total_output_tokens":3103340,"total_runtime_s":161452.09199999998,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.727232361384374,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c65a4361967bb6d6ff6d","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-858689101d1f123263f1bd4b","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.721190499337615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1210,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[604];split=unlicensed/near-term;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1210,"metadata":{"avgScore":0.721190499337615,"avg_score":0.721190499337615,"avg_score_percent":72.1190499337615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":137.01,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":321979548,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":4654794,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":224956.202,"source_flags":[],"source_harness":"claude_code","source_row_index":604,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":137.01,"totalDurationS":224956.202,"totalInputTokens":321979548,"totalOutputTokens":4654794,"total_cost_usd":137.01,"total_input_tokens":321979548,"total_output_tokens":4654794,"total_runtime_s":224956.202,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":72.1190499337615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-998c7250d826896b9a1c","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-7e357924fa2124b8d2c980bd","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.241291868854736,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":238,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[118];split=full/last-exam;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":238,"metadata":{"avgScore":0.241291868854736,"avg_score":0.241291868854736,"avg_score_percent":24.1291868854736,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":216.63,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":559312348,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":4782419,"passRate":0.10526315789473684,"pass_rate":0.10526315789473684,"pass_rate_percent":10.526315789473683,"passes":4,"runs":36,"runtime":344446.109,"source_flags":[],"source_harness":"claude_code","source_row_index":118,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":216.63,"totalDurationS":344446.109,"totalInputTokens":559312348,"totalOutputTokens":4782419,"total_cost_usd":216.63,"total_input_tokens":559312348,"total_output_tokens":4782419,"total_runtime_s":344446.109,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.1291868854736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e1ee75370d6027b691a0","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-726b505a0a230fe8f3ab309a","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5248561340297907,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":6,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[2];split=full;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":6,"metadata":{"avgScore":0.5248561340297907,"avg_score":0.5248561340297907,"avg_score_percent":52.48561340297907,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":486.47,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1234201310,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":12430863,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":148,"runtime":729633.2999999999,"source_flags":[],"source_harness":"claude_code","source_row_index":2,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":486.47,"totalDurationS":729633.2999999999,"totalInputTokens":1234201310,"totalOutputTokens":12430863,"total_cost_usd":486.47,"total_input_tokens":1234201310,"total_output_tokens":12430863,"total_runtime_s":729633.2999999999,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":52.48561340297907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d0dbbd0b28cacd098da1","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-57a444f2b20e7d8258b3de23","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.721190499337615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":374,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[186];split=full/near-term;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":374,"metadata":{"avgScore":0.721190499337615,"avg_score":0.721190499337615,"avg_score_percent":72.1190499337615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":137.01,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":321979548,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":4654794,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":224956.202,"source_flags":[],"source_harness":"claude_code","source_row_index":186,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":137.01,"totalDurationS":224956.202,"totalInputTokens":321979548,"totalOutputTokens":4654794,"total_cost_usd":137.01,"total_input_tokens":321979548,"total_output_tokens":4654794,"total_runtime_s":224956.202,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":72.1190499337615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5ef8d7bec342f056285e","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2786935ffa2a143b7451d10a","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.16919639939223963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":646,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[322];split=licensed;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":646,"metadata":{"avgScore":0.16919639939223963,"avg_score":0.16919639939223963,"avg_score_percent":16.919639939223963,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":256972501,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":1676047,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":106893.535,"source_flags":[],"source_harness":"claude_code","source_row_index":322,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":80.31,"totalDurationS":106893.535,"totalInputTokens":256972501,"totalOutputTokens":1676047,"total_cost_usd":80.31,"total_input_tokens":256972501,"total_output_tokens":1676047,"total_runtime_s":106893.535,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":16.919639939223963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-55589d0cc0dcdc7c8e03","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-1d514ce716421dba4ed8df81","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.28130858943822107,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1074,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[536];split=unlicensed/last-exam;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1074,"metadata":{"avgScore":0.28130858943822107,"avg_score":0.28130858943822107,"avg_score_percent":28.13085894382211,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":164.29,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":394942645,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":3581950,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":4,"runs":29,"runtime":260986.983,"source_flags":[],"source_harness":"claude_code","source_row_index":536,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":164.29,"totalDurationS":260986.983,"totalInputTokens":394942645,"totalOutputTokens":3581950,"total_cost_usd":164.29,"total_input_tokens":394942645,"total_output_tokens":3581950,"total_runtime_s":260986.983,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":28.13085894382211,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a498540357b424e6d124","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-13f70f68525e987dc3f9f5e1","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5526026381504506,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":840,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[419];split=unlicensed;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":840,"metadata":{"avgScore":0.5526026381504506,"avg_score":0.5526026381504506,"avg_score_percent":55.260263815045064,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":406.16,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":977228809,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":10754816,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":138,"runtime":622739.765,"source_flags":[],"source_harness":"claude_code","source_row_index":419,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":406.16,"totalDurationS":622739.765,"totalInputTokens":977228809,"totalOutputTokens":10754816,"total_cost_usd":406.16,"total_input_tokens":977228809,"total_output_tokens":10754816,"total_runtime_s":622739.765,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":55.260263815045064,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8ce418c309cb5d689c8a","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-0588d14ffb9c1eb5800e86cd","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.5112101968778014,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dccfd08cf97dbbba1183c6e30cf2bf97ba90b38b","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":722,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[360];split=linux_only;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":722,"metadata":{"avgScore":0.5112101968778014,"avg_score":0.5112101968778014,"avg_score_percent":51.12101968778015,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":249.16,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":559879430,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-3-8-max","output_tokens":6819773,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":102,"runtime":407974.578,"source_flags":[],"source_harness":"claude_code","source_row_index":360,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":249.16,"totalDurationS":407974.578,"totalInputTokens":559879430,"totalOutputTokens":6819773,"total_cost_usd":249.16,"total_input_tokens":559879430,"total_output_tokens":6819773,"total_runtime_s":407974.578,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.12101968778015,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7355bd572c00580b6baa","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ce445d401a68db0d5300aec4","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2907801418439716,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1311,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[655];split=unlicensed/overall;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1311,"metadata":{"avgScore":0.5526026381504506,"avg_score":0.5526026381504506,"avg_score_percent":55.260263815045064,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":406.16,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":977228809,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":10754816,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":138,"runtime":622739.765,"source_flags":[],"source_harness":"claude_code","source_row_index":655,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":406.16,"totalDurationS":622739.765,"totalInputTokens":977228809,"totalOutputTokens":10754816,"total_cost_usd":406.16,"total_input_tokens":977228809,"total_output_tokens":10754816,"total_runtime_s":622739.765,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.078014184397162,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-103b029e72d7ae4389cf","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c71929a5b39f7c2e03412568","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":477,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[238];split=full/overall;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":477,"metadata":{"avgScore":0.5248561340297907,"avg_score":0.5248561340297907,"avg_score_percent":52.48561340297907,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":486.47,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1234201310,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":12430863,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":148,"runtime":729633.2999999999,"source_flags":[],"source_harness":"claude_code","source_row_index":238,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":486.47,"totalDurationS":729633.2999999999,"totalInputTokens":1234201310,"totalOutputTokens":12430863,"total_cost_usd":486.47,"total_input_tokens":1234201310,"total_output_tokens":12430863,"total_runtime_s":729633.2999999999,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5ac2944b9de04a7837b9","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c45231103580b62a6aa1f2e9","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1209,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[604];split=unlicensed/near-term;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1209,"metadata":{"avgScore":0.721190499337615,"avg_score":0.721190499337615,"avg_score_percent":72.1190499337615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":137.01,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":321979548,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":4654794,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":224956.202,"source_flags":[],"source_harness":"claude_code","source_row_index":604,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":137.01,"totalDurationS":224956.202,"totalInputTokens":321979548,"totalOutputTokens":4654794,"total_cost_usd":137.01,"total_input_tokens":321979548,"total_output_tokens":4654794,"total_runtime_s":224956.202,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ad2e862376b7babb5ed0","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-b4b47586ba6276e049e54bc0","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.23636363636363636,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":123,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[61];split=full/full-spectrum;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":123,"metadata":{"avgScore":0.46354442255769973,"avg_score":0.46354442255769973,"avg_score_percent":46.354442255769975,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":149.3,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":392062495,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":3623554,"passRate":0.23636363636363636,"pass_rate":0.23636363636363636,"pass_rate_percent":23.636363636363637,"passes":13,"runs":53,"runtime":207159.591,"source_flags":[],"source_harness":"claude_code","source_row_index":61,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":149.3,"totalDurationS":207159.591,"totalInputTokens":392062495,"totalOutputTokens":3623554,"total_cost_usd":149.3,"total_input_tokens":392062495,"total_output_tokens":3623554,"total_runtime_s":207159.591,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":23.636363636363637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-57e8dcec05c1dea4f1ff","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-98e4700a8c50f9ea7f2bded0","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.26973684210526316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":5,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[2];split=full;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":5,"metadata":{"avgScore":0.5248561340297907,"avg_score":0.5248561340297907,"avg_score_percent":52.48561340297907,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":486.47,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":1234201310,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":12430863,"passRate":0.26973684210526316,"pass_rate":0.26973684210526316,"pass_rate_percent":26.973684210526315,"passes":41,"runs":148,"runtime":729633.2999999999,"source_flags":[],"source_harness":"claude_code","source_row_index":2,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":148,"tokens":null,"totalCostUsd":486.47,"totalDurationS":729633.2999999999,"totalInputTokens":1234201310,"totalOutputTokens":12430863,"total_cost_usd":486.47,"total_input_tokens":1234201310,"total_output_tokens":12430863,"total_runtime_s":729633.2999999999,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.973684210526315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-45d8d92e88784e16ad32","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-989dc70f1480a5d047ad67f5","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":645,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[322];split=licensed;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":645,"metadata":{"avgScore":0.16919639939223963,"avg_score":0.16919639939223963,"avg_score_percent":16.919639939223963,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":80.31,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":256972501,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":1676047,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":106893.535,"source_flags":[],"source_harness":"claude_code","source_row_index":322,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":10,"tokens":null,"totalCostUsd":80.31,"totalDurationS":106893.535,"totalInputTokens":256972501,"totalOutputTokens":1676047,"total_cost_usd":80.31,"total_input_tokens":256972501,"total_output_tokens":1676047,"total_runtime_s":106893.535,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-316c1bb87e28c20e68ba","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-41cbd96351d110d9513565d1","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.24,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":957,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[478];split=unlicensed/full-spectrum;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":957,"metadata":{"avgScore":0.46727232361384374,"avg_score":0.46727232361384374,"avg_score_percent":46.727232361384374,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":120.6,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":298336188,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":3103340,"passRate":0.24,"pass_rate":0.24,"pass_rate_percent":24.0,"passes":12,"runs":48,"runtime":161452.09199999998,"source_flags":[],"source_harness":"claude_code","source_row_index":478,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":48,"tokens":null,"totalCostUsd":120.6,"totalDurationS":161452.09199999998,"totalInputTokens":298336188,"totalOutputTokens":3103340,"total_cost_usd":120.6,"total_input_tokens":298336188,"total_output_tokens":3103340,"total_runtime_s":161452.09199999998,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-729097edcad6aa008a9b","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-2a295c4ccc1b71a4f4129c45","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.24761904761904763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":721,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[360];split=linux_only;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":721,"metadata":{"avgScore":0.5112101968778014,"avg_score":0.5112101968778014,"avg_score_percent":51.12101968778015,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":249.16,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":559879430,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":6819773,"passRate":0.24761904761904763,"pass_rate":0.24761904761904763,"pass_rate_percent":24.761904761904763,"passes":26,"runs":102,"runtime":407974.578,"source_flags":[],"source_harness":"claude_code","source_row_index":360,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":249.16,"totalDurationS":407974.578,"totalInputTokens":559879430,"totalOutputTokens":6819773,"total_cost_usd":249.16,"total_input_tokens":559879430,"total_output_tokens":6819773,"total_runtime_s":407974.578,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.761904761904763,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-41714907385857d645a6","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-285b3b0c506c72ce6f01c93e","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.13333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1073,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[536];split=unlicensed/last-exam;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1073,"metadata":{"avgScore":0.28130858943822107,"avg_score":0.28130858943822107,"avg_score_percent":28.13085894382211,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":164.29,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":394942645,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":3581950,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":4,"runs":29,"runtime":260986.983,"source_flags":[],"source_harness":"claude_code","source_row_index":536,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":164.29,"totalDurationS":260986.983,"totalInputTokens":394942645,"totalOutputTokens":3581950,"total_cost_usd":164.29,"total_input_tokens":394942645,"total_output_tokens":3581950,"total_runtime_s":260986.983,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":13.333333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-5c016cab515675898503","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-21c6cead33aa2af365038f15","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.2907801418439716,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":839,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[419];split=unlicensed;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":839,"metadata":{"avgScore":0.5526026381504506,"avg_score":0.5526026381504506,"avg_score_percent":55.260263815045064,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":406.16,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":977228809,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":10754816,"passRate":0.2907801418439716,"pass_rate":0.2907801418439716,"pass_rate_percent":29.078014184397162,"passes":41,"runs":138,"runtime":622739.765,"source_flags":[],"source_harness":"claude_code","source_row_index":419,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":138,"tokens":null,"totalCostUsd":406.16,"totalDurationS":622739.765,"totalInputTokens":977228809,"totalOutputTokens":10754816,"total_cost_usd":406.16,"total_input_tokens":977228809,"total_output_tokens":10754816,"total_runtime_s":622739.765,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.078014184397162,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9d59a9eed6911496e5d9","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-064e6e0ecae307f9da621cd4","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.10526315789473684,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":237,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[118];split=full/last-exam;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":237,"metadata":{"avgScore":0.241291868854736,"avg_score":0.241291868854736,"avg_score_percent":24.1291868854736,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":216.63,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":559312348,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":4782419,"passRate":0.10526315789473684,"pass_rate":0.10526315789473684,"pass_rate_percent":10.526315789473683,"passes":4,"runs":36,"runtime":344446.109,"source_flags":[],"source_harness":"claude_code","source_row_index":118,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":36,"tokens":null,"totalCostUsd":216.63,"totalDurationS":344446.109,"totalInputTokens":559312348,"totalOutputTokens":4782419,"total_cost_usd":216.63,"total_input_tokens":559312348,"total_output_tokens":4782419,"total_runtime_s":344446.109,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":10.526315789473683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9048760d2fd21b7f8e7f","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-00da1a25423a94340d8cba5e","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.8 Max","modelRef":"qwen-3-8-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"claude-code","harness_variant":"thinking-xhigh","source_harness":"claude_code","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","harness_specific","heuristic_model_mapping","missing_source_date","split_specific"],"rank":null,"rawValue":0.3880597014925373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4f4617ea7ea6dcaf9f1e8d74ec6f2814dea96e","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":373,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[186];split=full/near-term;harness=claude_code;model=qwen-3-8-max;variant=thinking-xhigh;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":373,"metadata":{"avgScore":0.721190499337615,"avg_score":0.721190499337615,"avg_score_percent":72.1190499337615,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":137.01,"costSource":"db","cost_source":"db","harness":"claude_code","harness_id":"claude-code","harness_variant":"thinking-xhigh","input_tokens":321979548,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-3-8-max","output_tokens":4654794,"passRate":0.3880597014925373,"pass_rate":0.3880597014925373,"pass_rate_percent":38.80597014925373,"passes":26,"runs":67,"runtime":224956.202,"source_flags":[],"source_harness":"claude_code","source_row_index":186,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":137.01,"totalDurationS":224956.202,"totalInputTokens":321979548,"totalOutputTokens":4654794,"total_cost_usd":137.01,"total_input_tokens":321979548,"total_output_tokens":4654794,"total_runtime_s":224956.202,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":38.80597014925373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a9f40ae16cf98706b5e6","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ef35a8b148e44b6a88a06eff","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3038085248805724,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1032,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[515];split=unlicensed/full-spectrum;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1032,"metadata":{"avgScore":0.3038085248805724,"avg_score":0.3038085248805724,"avg_score_percent":30.380852488057243,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":244.25,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":501226643,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":6740794,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":49,"runtime":93824,"source_flags":[],"source_harness":"openclaw","source_row_index":515,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":244.25,"totalDurationS":93824,"totalInputTokens":501226643,"totalOutputTokens":6740794,"total_cost_usd":244.25,"total_input_tokens":501226643,"total_output_tokens":6740794,"total_runtime_s":93824,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":30.380852488057243,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c20cf50f1b2364e6516","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c6600bdfb54d16080b6524f7","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.46895390009159155,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":450,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[224];split=full/near-term;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":450,"metadata":{"avgScore":0.46895390009159155,"avg_score":0.46895390009159155,"avg_score_percent":46.89539000915916,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.83,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":501983933,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":7323900,"passRate":0.1791044776119403,"pass_rate":0.1791044776119403,"pass_rate_percent":17.91044776119403,"passes":12,"runs":63,"runtime":137430,"source_flags":[],"source_harness":"openclaw","source_row_index":224,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":246.83,"totalDurationS":137430,"totalInputTokens":501983933,"totalOutputTokens":7323900,"total_cost_usd":246.83,"total_input_tokens":501983933,"total_output_tokens":7323900,"total_runtime_s":137430,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":46.89539000915916,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8a58dbda3e6e27fdae0b","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-c5d89e6fe75a21b3a1416df1","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.33691021666783877,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1396,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[697];split=unlicensed/overall;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1396,"metadata":{"avgScore":0.33691021666783877,"avg_score":0.33691021666783877,"avg_score_percent":33.691021666783875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":603.35,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1245322848,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":17017497,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":135,"runtime":312987,"source_flags":[],"source_harness":"openclaw","source_row_index":697,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":603.35,"totalDurationS":312987,"totalInputTokens":1245322848,"totalOutputTokens":17017497,"total_cost_usd":603.35,"total_input_tokens":1245322848,"total_output_tokens":17017497,"total_runtime_s":312987,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.691021666783875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-3c2f842bb439ca366fc7","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b5f57d56beb206f05cc9534e","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2849168408005203,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":202,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[100];split=full/full-spectrum;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":202,"metadata":{"avgScore":0.2849168408005203,"avg_score":0.2849168408005203,"avg_score_percent":28.491684080052032,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":280.32,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":585062442,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":7084714,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":53,"runtime":128744,"source_flags":[],"source_harness":"openclaw","source_row_index":100,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":280.32,"totalDurationS":128744,"totalInputTokens":585062442,"totalOutputTokens":7084714,"total_cost_usd":280.32,"total_input_tokens":585062442,"total_output_tokens":7084714,"total_runtime_s":128744,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.491684080052032,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9604794add9a46c17835","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9c56c39fc1c43c2a7449bf60","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.34638125058546504,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":798,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[398];split=linux_only;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":798,"metadata":{"avgScore":0.34638125058546504,"avg_score":0.34638125058546504,"avg_score_percent":34.63812505854651,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":534.27,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1014954271,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":15755710,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":14,"runs":102,"runtime":123757,"source_flags":[],"source_harness":"openclaw","source_row_index":398,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":534.27,"totalDurationS":123757,"totalInputTokens":1014954271,"totalOutputTokens":15755710,"total_cost_usd":534.27,"total_input_tokens":1014954271,"total_output_tokens":15755710,"total_runtime_s":123757,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":34.63812505854651,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-656301f5ed106a1e8e2a","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-9c2311cf43d0d3869fdf8fa9","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3156864509879294,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":564,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[281];split=full/overall;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":564,"metadata":{"avgScore":0.3156864509879294,"avg_score":0.3156864509879294,"avg_score_percent":31.568645098792942,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":663.87,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1382717830,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":17557559,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":142,"runtime":363579,"source_flags":[],"source_harness":"openclaw","source_row_index":281,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":663.87,"totalDurationS":363579,"totalInputTokens":1382717830,"totalOutputTokens":17557559,"total_cost_usd":663.87,"total_input_tokens":1382717830,"total_output_tokens":17557559,"total_runtime_s":363579,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.568645098792942,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7f553ff3b39d48339589","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-862ba707b7e9473167773ac1","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.33691021666783877,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":924,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[461];split=unlicensed;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":924,"metadata":{"avgScore":0.33691021666783877,"avg_score":0.33691021666783877,"avg_score_percent":33.691021666783875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":603.35,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1245322848,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":17017497,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":135,"runtime":312987,"source_flags":[],"source_harness":"openclaw","source_row_index":461,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":603.35,"totalDurationS":312987,"totalInputTokens":1245322848,"totalOutputTokens":17017497,"total_cost_usd":603.35,"total_input_tokens":1245322848,"total_output_tokens":17017497,"total_runtime_s":312987,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":33.691021666783875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7f703fb191a408bfa21b","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6320bfc742dd7ee25123a2e0","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06390764530880215,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":324,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[161];split=full/last-exam;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":324,"metadata":{"avgScore":0.06390764530880215,"avg_score":0.06390764530880215,"avg_score_percent":6.390764530880215,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":173.41,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":377757993,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":4129312,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":33,"runtime":118997,"source_flags":[],"source_harness":"openclaw","source_row_index":161,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":173.41,"totalDurationS":118997,"totalInputTokens":377757993,"totalOutputTokens":4129312,"total_cost_usd":173.41,"total_input_tokens":377757993,"total_output_tokens":4129312,"total_runtime_s":118997,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.390764530880215,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d1daac2d1aac58bee905","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-364c44e9b8f56bb6e6c9d6b3","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.3156864509879294,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":92,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[45];split=full;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":92,"metadata":{"avgScore":0.3156864509879294,"avg_score":0.3156864509879294,"avg_score_percent":31.568645098792942,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":663.87,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1382717830,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":17557559,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":142,"runtime":363579,"source_flags":[],"source_harness":"openclaw","source_row_index":45,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":663.87,"totalDurationS":363579,"totalInputTokens":1382717830,"totalOutputTokens":17557559,"total_cost_usd":663.87,"total_input_tokens":1382717830,"total_output_tokens":17557559,"total_runtime_s":363579,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":31.568645098792942,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-da8e2859dfea79cb9b8e","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1bf3a7f7f5905fb899478b3e","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.04363636363636363,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":696,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[347];split=licensed;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":696,"metadata":{"avgScore":0.04363636363636363,"avg_score":0.04363636363636363,"avg_score_percent":4.363636363636363,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":60.52,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":137394982,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":540062,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":50592,"source_flags":[],"source_harness":"openclaw","source_row_index":347,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":60.52,"totalDurationS":50592,"totalInputTokens":137394982,"totalOutputTokens":540062,"total_cost_usd":60.52,"total_input_tokens":137394982,"total_output_tokens":540062,"total_runtime_s":50592,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.363636363636363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d8bc9174b5f1aa5031f4","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0d4c9fc6ca0891fd6935ef1e","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.46895390009159155,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1286,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[642];split=unlicensed/near-term;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1286,"metadata":{"avgScore":0.46895390009159155,"avg_score":0.46895390009159155,"avg_score_percent":46.89539000915916,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.83,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":501983933,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":7323900,"passRate":0.1791044776119403,"pass_rate":0.1791044776119403,"pass_rate_percent":17.91044776119403,"passes":12,"runs":63,"runtime":137430,"source_flags":[],"source_harness":"openclaw","source_row_index":642,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":246.83,"totalDurationS":137430,"totalInputTokens":501983933,"totalOutputTokens":7323900,"total_cost_usd":246.83,"total_input_tokens":501983933,"total_output_tokens":7323900,"total_runtime_s":137430,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":46.89539000915916,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8ea168637e38b085225c","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0bab71f706949de7a145f5b7","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08094968405781607,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c909e375e2921155cc530c9feb46ef94d0096f26","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1156,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[577];split=unlicensed/last-exam;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1156,"metadata":{"avgScore":0.08094968405781607,"avg_score":0.08094968405781607,"avg_score_percent":8.094968405781607,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":136.99,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":291906865,"metric_id":"avg_score","metric_role":"secondary","model":"qwen-qwen3-7-max","output_tokens":3821425,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":29,"runtime":95778,"source_flags":[],"source_harness":"openclaw","source_row_index":577,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":136.99,"totalDurationS":95778,"totalInputTokens":291906865,"totalOutputTokens":3821425,"total_cost_usd":136.99,"total_input_tokens":291906865,"total_output_tokens":3821425,"total_runtime_s":95778,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.094968405781607,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4ec1279086c517ec86f5","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-e35ea090a2fe30f23a25fca5","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.13333333333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":797,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[398];split=linux_only;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":797,"metadata":{"avgScore":0.34638125058546504,"avg_score":0.34638125058546504,"avg_score_percent":34.63812505854651,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":534.27,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1014954271,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":15755710,"passRate":0.13333333333333333,"pass_rate":0.13333333333333333,"pass_rate_percent":13.333333333333334,"passes":14,"runs":102,"runtime":123757,"source_flags":[],"source_harness":"openclaw","source_row_index":398,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":102,"tokens":null,"totalCostUsd":534.27,"totalDurationS":123757,"totalInputTokens":1014954271,"totalOutputTokens":15755710,"total_cost_usd":534.27,"total_input_tokens":1014954271,"total_output_tokens":15755710,"total_runtime_s":123757,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":13.333333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-205d4665334cf62d8202","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-cba7c708dd10cb5ae6c36463","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":323,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[161];split=full/last-exam;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":323,"metadata":{"avgScore":0.06390764530880215,"avg_score":0.06390764530880215,"avg_score_percent":6.390764530880215,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":173.41,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":377757993,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":4129312,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":33,"runtime":118997,"source_flags":[],"source_harness":"openclaw","source_row_index":161,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":173.41,"totalDurationS":118997,"totalInputTokens":377757993,"totalOutputTokens":4129312,"total_cost_usd":173.41,"total_input_tokens":377757993,"total_output_tokens":4129312,"total_runtime_s":118997,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4dfc6215f55ab8e29c65","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-bcf905e6f892904ae60d1f19","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1155,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[577];split=unlicensed/last-exam;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1155,"metadata":{"avgScore":0.08094968405781607,"avg_score":0.08094968405781607,"avg_score_percent":8.094968405781607,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":136.99,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":291906865,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":3821425,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":29,"runtime":95778,"source_flags":[],"source_harness":"openclaw","source_row_index":577,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":136.99,"totalDurationS":95778,"totalInputTokens":291906865,"totalOutputTokens":3821425,"total_cost_usd":136.99,"total_input_tokens":291906865,"total_output_tokens":3821425,"total_runtime_s":95778,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-c8ee607107ea8c1dca71","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-aca635ee1d6d752cbf6e7b48","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11842105263157894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":91,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[45];split=full;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":91,"metadata":{"avgScore":0.3156864509879294,"avg_score":0.3156864509879294,"avg_score_percent":31.568645098792942,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":663.87,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1382717830,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":17557559,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":142,"runtime":363579,"source_flags":[],"source_harness":"openclaw","source_row_index":45,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":663.87,"totalDurationS":363579,"totalInputTokens":1382717830,"totalOutputTokens":17557559,"total_cost_usd":663.87,"total_input_tokens":1382717830,"total_output_tokens":17557559,"total_runtime_s":363579,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.842105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ed6d02090ed9dbc1de89","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-8c9f41edb00afb8498dccad0","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1276595744680851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1395,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[697];split=unlicensed/overall;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1395,"metadata":{"avgScore":0.33691021666783877,"avg_score":0.33691021666783877,"avg_score_percent":33.691021666783875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":603.35,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1245322848,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":17017497,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":135,"runtime":312987,"source_flags":[],"source_harness":"openclaw","source_row_index":697,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":603.35,"totalDurationS":312987,"totalInputTokens":1245322848,"totalOutputTokens":17017497,"total_cost_usd":603.35,"total_input_tokens":1245322848,"total_output_tokens":17017497,"total_runtime_s":312987,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.76595744680851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a30f59bc00e907f674a5","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-7baaed58e866c9c766d9b134","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.11842105263157894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":563,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[281];split=full/overall;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":563,"metadata":{"avgScore":0.3156864509879294,"avg_score":0.3156864509879294,"avg_score_percent":31.568645098792942,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":663.87,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1382717830,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":17557559,"passRate":0.11842105263157894,"pass_rate":0.11842105263157894,"pass_rate_percent":11.842105263157894,"passes":18,"runs":142,"runtime":363579,"source_flags":[],"source_harness":"openclaw","source_row_index":281,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":142,"tokens":null,"totalCostUsd":663.87,"totalDurationS":363579,"totalInputTokens":1382717830,"totalOutputTokens":17557559,"total_cost_usd":663.87,"total_input_tokens":1382717830,"total_output_tokens":17557559,"total_runtime_s":363579,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":11.842105263157894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d77da36d3252c147ce78","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6d8cf11c2f50d09190c0e21b","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1276595744680851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":923,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[461];split=unlicensed;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":923,"metadata":{"avgScore":0.33691021666783877,"avg_score":0.33691021666783877,"avg_score_percent":33.691021666783875,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":603.35,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":1245322848,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":17017497,"passRate":0.1276595744680851,"pass_rate":0.1276595744680851,"pass_rate_percent":12.76595744680851,"passes":18,"runs":135,"runtime":312987,"source_flags":[],"source_harness":"openclaw","source_row_index":461,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":135,"tokens":null,"totalCostUsd":603.35,"totalDurationS":312987,"totalInputTokens":1245322848,"totalOutputTokens":17017497,"total_cost_usd":603.35,"total_input_tokens":1245322848,"total_output_tokens":17017497,"total_runtime_s":312987,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.76595744680851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-e5e6efd708ba2b99be9b","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-571f562ea42434821e9f1a7a","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1031,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[515];split=unlicensed/full-spectrum;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1031,"metadata":{"avgScore":0.3038085248805724,"avg_score":0.3038085248805724,"avg_score_percent":30.380852488057243,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":244.25,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":501226643,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":6740794,"passRate":0.12,"pass_rate":0.12,"pass_rate_percent":12.0,"passes":6,"runs":49,"runtime":93824,"source_flags":[],"source_harness":"openclaw","source_row_index":515,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":244.25,"totalDurationS":93824,"totalInputTokens":501226643,"totalOutputTokens":6740794,"total_cost_usd":244.25,"total_input_tokens":501226643,"total_output_tokens":6740794,"total_runtime_s":93824,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-8d3b1daa733f4dd95e2d","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-3efc3f6f42d3bbbefdf1aa55","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1791044776119403,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":449,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[224];split=full/near-term;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":449,"metadata":{"avgScore":0.46895390009159155,"avg_score":0.46895390009159155,"avg_score_percent":46.89539000915916,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.83,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":501983933,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":7323900,"passRate":0.1791044776119403,"pass_rate":0.1791044776119403,"pass_rate_percent":17.91044776119403,"passes":12,"runs":63,"runtime":137430,"source_flags":[],"source_harness":"openclaw","source_row_index":224,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":246.83,"totalDurationS":137430,"totalInputTokens":501983933,"totalOutputTokens":7323900,"total_cost_usd":246.83,"total_input_tokens":501983933,"total_output_tokens":7323900,"total_runtime_s":137430,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.91044776119403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-9c661e8110bcfa4a43b8","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-396f4ab7cf6946f2ad6ab4d6","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.1791044776119403,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1285,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[642];split=unlicensed/near-term;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1285,"metadata":{"avgScore":0.46895390009159155,"avg_score":0.46895390009159155,"avg_score_percent":46.89539000915916,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":246.83,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":501983933,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":7323900,"passRate":0.1791044776119403,"pass_rate":0.1791044776119403,"pass_rate_percent":17.91044776119403,"passes":12,"runs":63,"runtime":137430,"source_flags":[],"source_harness":"openclaw","source_row_index":642,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":63,"tokens":null,"totalCostUsd":246.83,"totalDurationS":137430,"totalInputTokens":501983933,"totalOutputTokens":7323900,"total_cost_usd":246.83,"total_input_tokens":501983933,"total_output_tokens":7323900,"total_runtime_s":137430,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":17.91044776119403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-873e607a1cd86eb4a4c1","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-24710b71792065fbc6a83b78","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10909090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":201,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[100];split=full/full-spectrum;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":201,"metadata":{"avgScore":0.2849168408005203,"avg_score":0.2849168408005203,"avg_score_percent":28.491684080052032,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":280.32,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":585062442,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":7084714,"passRate":0.10909090909090909,"pass_rate":0.10909090909090909,"pass_rate_percent":10.909090909090908,"passes":6,"runs":53,"runtime":128744,"source_flags":[],"source_harness":"openclaw","source_row_index":100,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":53,"tokens":null,"totalCostUsd":280.32,"totalDurationS":128744,"totalInputTokens":585062442,"totalOutputTokens":7084714,"total_cost_usd":280.32,"total_input_tokens":585062442,"total_output_tokens":7084714,"total_runtime_s":128744,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.909090909090908,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4b264ab384418b9118d5","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1267f5558705fbdc526b6adc","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","registry":"model_aliases.json"},"mappingNote":"Agents' Last Exam spelling of the explicit Qwen3.7 Max release; harness remains attached to the system row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.7 Max","modelRef":"qwen-qwen3-7-max","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db+list","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80fb5285d01d85a97a456283f3c955e0b26410d4","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":695,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[347];split=licensed;harness=openclaw;model=qwen-qwen3-7-max;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":695,"metadata":{"avgScore":0.04363636363636363,"avg_score":0.04363636363636363,"avg_score_percent":4.363636363636363,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":60.52,"costSource":"db+list","cost_source":"db+list","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":137394982,"metric_id":"pass_rate","metric_role":"primary","model":"qwen-qwen3-7-max","output_tokens":540062,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":7,"runtime":50592,"source_flags":[],"source_harness":"openclaw","source_row_index":347,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":60.52,"totalDurationS":50592,"totalInputTokens":137394982,"totalOutputTokens":540062,"total_cost_usd":60.52,"total_input_tokens":137394982,"total_output_tokens":540062,"total_runtime_s":50592,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f67d182a48f56d4141cb","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-f6767105d6b9b794e702fee2","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26021994074929344,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1412,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[705];split=unlicensed/overall;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1412,"metadata":{"avgScore":0.26021994074929344,"avg_score":0.26021994074929344,"avg_score_percent":26.021994074929346,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":213.49,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":618181138.6666666,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":6451781.833333333,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":262,"runtime":438548.9462786893,"source_flags":[],"source_harness":"openclaw","source_row_index":705,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":213.49,"totalDurationS":438548.9462786893,"totalInputTokens":618181138.6666666,"totalOutputTokens":6451781.833333333,"total_cost_usd":213.49,"total_input_tokens":618181138.6666666,"total_output_tokens":6451781.833333333,"total_runtime_s":438548.9462786893,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.021994074929346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-86dd928227f047ba5fbf","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b85bfce593ecc0e6cfbce742","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.06469090909090909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":680,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[339];split=licensed;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":680,"metadata":{"avgScore":0.06469090909090909,"avg_score":0.06469090909090909,"avg_score_percent":6.469090909090909,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":40.36,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":120729225.5,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":575606.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":46421.8629133185,"source_flags":[],"source_harness":"openclaw","source_row_index":339,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":40.36,"totalDurationS":46421.8629133185,"totalInputTokens":120729225.5,"totalOutputTokens":575606.5,"total_cost_usd":40.36,"total_input_tokens":120729225.5,"total_output_tokens":575606.5,"total_runtime_s":46421.8629133185,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":6.469090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0d66acfda6d3da1dcad4","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-af2119930547288707824046","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.26021994074929344,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":940,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[469];split=unlicensed;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":940,"metadata":{"avgScore":0.26021994074929344,"avg_score":0.26021994074929344,"avg_score_percent":26.021994074929346,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":213.49,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":618181138.6666666,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":6451781.833333333,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":262,"runtime":438548.9462786893,"source_flags":[],"source_harness":"openclaw","source_row_index":469,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":213.49,"totalDurationS":438548.9462786893,"totalInputTokens":618181138.6666666,"totalOutputTokens":6451781.833333333,"total_cost_usd":213.49,"total_input_tokens":618181138.6666666,"total_output_tokens":6451781.833333333,"total_runtime_s":438548.9462786893,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":26.021994074929346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6eed8f7e4cedc1ca6d04","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ab5cf659a72531bf2d122930","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2895843637929504,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":820,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[409];split=linux_only;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":820,"metadata":{"avgScore":0.2895843637929504,"avg_score":0.2895843637929504,"avg_score_percent":28.958436379295037,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.87,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":366815218.1666667,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":4436483.5,"passRate":0.10476190476190476,"pass_rate":0.10476190476190476,"pass_rate_percent":10.476190476190476,"passes":21,"runs":201,"runtime":224862.70664021675,"source_flags":[],"source_harness":"openclaw","source_row_index":409,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":127.87,"totalDurationS":224862.70664021675,"totalInputTokens":366815218.1666667,"totalOutputTokens":4436483.5,"total_cost_usd":127.87,"total_input_tokens":366815218.1666667,"total_output_tokens":4436483.5,"total_runtime_s":224862.70664021675,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":28.958436379295037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-33ddc2d44843e03a59e8","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-a9c43120c5f0c00370db8db5","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.2290770537090957,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":218,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[108];split=full/full-spectrum;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":218,"metadata":{"avgScore":0.2290770537090957,"avg_score":0.2290770537090957,"avg_score_percent":22.90770537090957,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.66,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":322063737.3333333,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":3070232.3333333335,"passRate":0.08181818181818182,"pass_rate":0.08181818181818182,"pass_rate_percent":8.181818181818182,"passes":8,"runs":97,"runtime":167817.14606574166,"source_flags":[],"source_harness":"openclaw","source_row_index":108,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":110.66,"totalDurationS":167817.14606574166,"totalInputTokens":322063737.3333333,"totalOutputTokens":3070232.3333333335,"total_cost_usd":110.66,"total_input_tokens":322063737.3333333,"total_output_tokens":3070232.3333333335,"total_runtime_s":167817.14606574166,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":22.90770537090957,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-712ab0b2dc2c824ff5ec","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-78980a33acc55dc5bafb2ccf","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.36422336106940467,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":458,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[228];split=full/near-term;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":458,"metadata":{"avgScore":0.36422336106940467,"avg_score":0.36422336106940467,"avg_score_percent":36.42233610694047,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.59,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":259732239.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":2654193.3333333335,"passRate":0.12686567164179105,"pass_rate":0.12686567164179105,"pass_rate_percent":12.686567164179104,"passes":17,"runs":131,"runtime":216764.1855777436,"source_flags":[],"source_harness":"openclaw","source_row_index":228,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":89.59,"totalDurationS":216764.1855777436,"totalInputTokens":259732239.16666666,"totalOutputTokens":2654193.3333333335,"total_cost_usd":89.59,"total_input_tokens":259732239.16666666,"total_output_tokens":2654193.3333333335,"total_runtime_s":216764.1855777436,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":36.42233610694047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4a2b2c774a21235cb2e0","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-6637535f1ea23fd95083e30e","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24606981345822615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":104,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[51];split=full;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":104,"metadata":{"avgScore":0.24606981345822615,"avg_score":0.24606981345822615,"avg_score_percent":24.606981345822614,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":253.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":738910364.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":7027388.333333333,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":272,"runtime":484970.8091920079,"source_flags":[],"source_harness":"openclaw","source_row_index":51,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":253.85,"totalDurationS":484970.8091920079,"totalInputTokens":738910364.1666666,"totalOutputTokens":7027388.333333333,"total_cost_usd":253.85,"total_input_tokens":738910364.1666666,"total_output_tokens":7027388.333333333,"total_runtime_s":484970.8091920079,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.606981345822614,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-6d6dc3360475c9ac3679","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-4bd6d899b27afa77a6a46d69","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.0464254429111199,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1168,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[583];split=unlicensed/last-exam;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1168,"metadata":{"avgScore":0.0464254429111199,"avg_score":0.0464254429111199,"avg_score_percent":4.64254429111199,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.08,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":177691032.83333334,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":1710081.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":54,"runtime":115560.64449874574,"source_flags":[],"source_harness":"openclaw","source_row_index":583,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":61.08,"totalDurationS":115560.64449874574,"totalInputTokens":177691032.83333334,"totalOutputTokens":1710081.8333333333,"total_cost_usd":61.08,"total_input_tokens":177691032.83333334,"total_output_tokens":1710081.8333333333,"total_runtime_s":115560.64449874574,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.64254429111199,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-ebc09cf1630f2db484d5","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-45dd52ab47322da58dc547a3","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.36422336106940467,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1294,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[646];split=unlicensed/near-term;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1294,"metadata":{"avgScore":0.36422336106940467,"avg_score":0.36422336106940467,"avg_score_percent":36.42233610694047,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.59,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":259732239.16666666,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":2654193.3333333335,"passRate":0.12686567164179105,"pass_rate":0.12686567164179105,"pass_rate_percent":12.686567164179104,"passes":17,"runs":131,"runtime":216764.1855777436,"source_flags":[],"source_harness":"openclaw","source_row_index":646,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":89.59,"totalDurationS":216764.1855777436,"totalInputTokens":259732239.16666666,"totalOutputTokens":2654193.3333333335,"total_cost_usd":89.59,"total_input_tokens":259732239.16666666,"total_output_tokens":2654193.3333333335,"total_runtime_s":216764.1855777436,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":36.42233610694047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cb2fba44de3b3c2e9dcc","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2f7910b15f1ad5068ab27035","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24758475908000527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1048,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[523];split=unlicensed/full-spectrum;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1048,"metadata":{"avgScore":0.24758475908000527,"avg_score":0.24758475908000527,"avg_score_percent":24.758475908000527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.18,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":228166672.83333334,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":2579788.3333333335,"passRate":0.09,"pass_rate":0.09,"pass_rate_percent":9.0,"passes":8,"runs":89,"runtime":127467.2211974289,"source_flags":[],"source_harness":"openclaw","source_row_index":523,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":79.18,"totalDurationS":127467.2211974289,"totalInputTokens":228166672.83333334,"totalOutputTokens":2579788.3333333335,"total_cost_usd":79.18,"total_input_tokens":228166672.83333334,"total_output_tokens":2579788.3333333335,"total_runtime_s":127467.2211974289,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.758475908000527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-52a3adcb7731260e8acc","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-1edbb3b9b7efed208ad28d7f","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.049588507561410436,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":332,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[165];split=full/last-exam;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":332,"metadata":{"avgScore":0.049588507561410436,"avg_score":0.049588507561410436,"avg_score_percent":4.958850756141044,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":78.19,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":228904638.83333334,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":1944406.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":59,"runtime":128733.20068769388,"source_flags":[],"source_harness":"openclaw","source_row_index":165,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":78.19,"totalDurationS":128733.20068769388,"totalInputTokens":228904638.83333334,"totalOutputTokens":1944406.8333333333,"total_cost_usd":78.19,"total_input_tokens":228904638.83333334,"total_output_tokens":1944406.8333333333,"total_runtime_s":128733.20068769388,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":4.958850756141044,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4433957ad2dfdd5c1552","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-07ed112c4808f34f7960a317","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"avg_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.24606981345822615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46a6373d5bc0cf3ce74f5b14cab43451a53a5f2d","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":576,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[287];split=full/overall;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=avg_score","sourceMetricId":"avg_score","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":576,"metadata":{"avgScore":0.24606981345822615,"avg_score":0.24606981345822615,"avg_score_percent":24.606981345822614,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":253.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":738910364.1666666,"metric_id":"avg_score","metric_role":"secondary","model":"qwen3-6-plus","output_tokens":7027388.333333333,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":272,"runtime":484970.8091920079,"source_flags":[],"source_harness":"openclaw","source_row_index":287,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":253.85,"totalDurationS":484970.8091920079,"totalInputTokens":738910364.1666666,"totalOutputTokens":7027388.333333333,"total_cost_usd":253.85,"total_input_tokens":738910364.1666666,"total_output_tokens":7027388.333333333,"total_runtime_s":484970.8091920079,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":24.606981345822614,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-cd132698f960883addb4","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-ecb5d7bdd21f4f243825c784","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/overall","split_benchmark_ref":"agents-last-exam-v1-full-overall","split_track":"full/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08552631578947369,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":575,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[287];split=full/overall;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":575,"metadata":{"avgScore":0.24606981345822615,"avg_score":0.24606981345822615,"avg_score_percent":24.606981345822614,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":253.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":738910364.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":7027388.333333333,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":272,"runtime":484970.8091920079,"source_flags":[],"source_harness":"openclaw","source_row_index":287,"source_status":"official_api_reported","split":"full/overall","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full-overall","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":253.85,"totalDurationS":484970.8091920079,"totalInputTokens":738910364.1666666,"totalOutputTokens":7027388.333333333,"total_cost_usd":253.85,"total_input_tokens":738910364.1666666,"total_output_tokens":7027388.333333333,"total_runtime_s":484970.8091920079,"total_tokens":null,"track":"full/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.552631578947368,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-4436032ad9d0b75095ab","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-db083475472d6c119d877ecf","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_track":"full/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08181818181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":217,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[108];split=full/full-spectrum;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":217,"metadata":{"avgScore":0.2290770537090957,"avg_score":0.2290770537090957,"avg_score_percent":22.90770537090957,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":110.66,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":322063737.3333333,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":3070232.3333333335,"passRate":0.08181818181818182,"pass_rate":0.08181818181818182,"pass_rate_percent":8.181818181818182,"passes":8,"runs":97,"runtime":167817.14606574166,"source_flags":[],"source_harness":"openclaw","source_row_index":108,"source_status":"official_api_reported","split":"full/full-spectrum","splitTasks":55,"split_benchmark_ref":"agents-last-exam-v1-full-full-spectrum","split_tasks":55,"tasks":54,"tokens":null,"totalCostUsd":110.66,"totalDurationS":167817.14606574166,"totalInputTokens":322063737.3333333,"totalOutputTokens":3070232.3333333335,"total_cost_usd":110.66,"total_input_tokens":322063737.3333333,"total_output_tokens":3070232.3333333335,"total_runtime_s":167817.14606574166,"total_tokens":null,"track":"full/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.181818181818182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-076e5c3d15cd131822bd","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-d812c68514c1aaacf7cbfd42","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/overall","split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_track":"unlicensed/overall","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09219858156028368,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1411,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[705];split=unlicensed/overall;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1411,"metadata":{"avgScore":0.26021994074929344,"avg_score":0.26021994074929344,"avg_score_percent":26.021994074929346,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":213.49,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":618181138.6666666,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":6451781.833333333,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":262,"runtime":438548.9462786893,"source_flags":[],"source_harness":"openclaw","source_row_index":705,"source_status":"official_api_reported","split":"unlicensed/overall","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-overall","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":213.49,"totalDurationS":438548.9462786893,"totalInputTokens":618181138.6666666,"totalOutputTokens":6451781.833333333,"total_cost_usd":213.49,"total_input_tokens":618181138.6666666,"total_output_tokens":6451781.833333333,"total_runtime_s":438548.9462786893,"total_tokens":null,"track":"unlicensed/overall"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.219858156028367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-42863bcb1b599db70507","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b811205a055192ffb17be886","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/last-exam","split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_track":"full/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":331,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[165];split=full/last-exam;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":331,"metadata":{"avgScore":0.049588507561410436,"avg_score":0.049588507561410436,"avg_score_percent":4.958850756141044,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":78.19,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":228904638.83333334,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":1944406.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":59,"runtime":128733.20068769388,"source_flags":[],"source_harness":"openclaw","source_row_index":165,"source_status":"official_api_reported","split":"full/last-exam","splitTasks":38,"split_benchmark_ref":"agents-last-exam-v1-full-last-exam","split_tasks":38,"tasks":33,"tokens":null,"totalCostUsd":78.19,"totalDurationS":128733.20068769388,"totalInputTokens":228904638.83333334,"totalOutputTokens":1944406.8333333333,"total_cost_usd":78.19,"total_input_tokens":228904638.83333334,"total_output_tokens":1944406.8333333333,"total_runtime_s":128733.20068769388,"total_tokens":null,"track":"full/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a78f43c900847f8faa8d","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-b7f9e814a9267c60d5f0c917","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full/near-term","split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_track":"full/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12686567164179105,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":457,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[228];split=full/near-term;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":457,"metadata":{"avgScore":0.36422336106940467,"avg_score":0.36422336106940467,"avg_score_percent":36.42233610694047,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.59,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":259732239.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":2654193.3333333335,"passRate":0.12686567164179105,"pass_rate":0.12686567164179105,"pass_rate_percent":12.686567164179104,"passes":17,"runs":131,"runtime":216764.1855777436,"source_flags":[],"source_harness":"openclaw","source_row_index":228,"source_status":"official_api_reported","split":"full/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-full-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":89.59,"totalDurationS":216764.1855777436,"totalInputTokens":259732239.16666666,"totalOutputTokens":2654193.3333333335,"total_cost_usd":89.59,"total_input_tokens":259732239.16666666,"total_output_tokens":2654193.3333333335,"total_runtime_s":216764.1855777436,"total_tokens":null,"track":"full/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.686567164179104,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-f190cdbfd946c6898dbc","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-83c9939c9cbe6695c45042a7","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/last-exam","split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_track":"unlicensed/last-exam","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1167,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[583];split=unlicensed/last-exam;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1167,"metadata":{"avgScore":0.0464254429111199,"avg_score":0.0464254429111199,"avg_score_percent":4.64254429111199,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":61.08,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":177691032.83333334,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":1710081.8333333333,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":54,"runtime":115560.64449874574,"source_flags":[],"source_harness":"openclaw","source_row_index":583,"source_status":"official_api_reported","split":"unlicensed/last-exam","splitTasks":30,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-last-exam","split_tasks":30,"tasks":29,"tokens":null,"totalCostUsd":61.08,"totalDurationS":115560.64449874574,"totalInputTokens":177691032.83333334,"totalOutputTokens":1710081.8333333333,"total_cost_usd":61.08,"total_input_tokens":177691032.83333334,"total_output_tokens":1710081.8333333333,"total_runtime_s":115560.64449874574,"total_tokens":null,"track":"unlicensed/last-exam"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-0e7a5be9d0b90c4b83b8","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-642b38a482b23a058bf45c0e","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/full-spectrum","split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_track":"unlicensed/full-spectrum","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1047,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[523];split=unlicensed/full-spectrum;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1047,"metadata":{"avgScore":0.24758475908000527,"avg_score":0.24758475908000527,"avg_score_percent":24.758475908000527,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":79.18,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":228166672.83333334,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":2579788.3333333335,"passRate":0.09,"pass_rate":0.09,"pass_rate_percent":9.0,"passes":8,"runs":89,"runtime":127467.2211974289,"source_flags":[],"source_harness":"openclaw","source_row_index":523,"source_status":"official_api_reported","split":"unlicensed/full-spectrum","splitTasks":50,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-full-spectrum","split_tasks":50,"tasks":49,"tokens":null,"totalCostUsd":79.18,"totalDurationS":127467.2211974289,"totalInputTokens":228166672.83333334,"totalOutputTokens":2579788.3333333335,"total_cost_usd":79.18,"total_input_tokens":228166672.83333334,"total_output_tokens":2579788.3333333335,"total_runtime_s":127467.2211974289,"total_tokens":null,"track":"unlicensed/full-spectrum"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d7cc5e1ae0cb49e7e484","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-5e3d409d8c25b727f1b9106b","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"full","split_benchmark_ref":"agents-last-exam-v1-full","split_track":"full","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.08552631578947369,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":103,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[51];split=full;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":103,"metadata":{"avgScore":0.24606981345822615,"avg_score":0.24606981345822615,"avg_score_percent":24.606981345822614,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":253.85,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":738910364.1666666,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":7027388.333333333,"passRate":0.08552631578947369,"pass_rate":0.08552631578947369,"pass_rate_percent":8.552631578947368,"passes":25,"runs":272,"runtime":484970.8091920079,"source_flags":[],"source_harness":"openclaw","source_row_index":51,"source_status":"official_api_reported","split":"full","splitTasks":152,"split_benchmark_ref":"agents-last-exam-v1-full","split_tasks":152,"tasks":146,"tokens":null,"totalCostUsd":253.85,"totalDurationS":484970.8091920079,"totalInputTokens":738910364.1666666,"totalOutputTokens":7027388.333333333,"total_cost_usd":253.85,"total_input_tokens":738910364.1666666,"total_output_tokens":7027388.333333333,"total_runtime_s":484970.8091920079,"total_tokens":null,"track":"full"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":8.552631578947368,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-7452912f4404b530af1a","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-39d39fbee2144380cf676de4","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"linux_only","split_benchmark_ref":"agents-last-exam-v1-linux-only","split_track":"linux_only","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.10476190476190476,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":819,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[409];split=linux_only;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":819,"metadata":{"avgScore":0.2895843637929504,"avg_score":0.2895843637929504,"avg_score_percent":28.958436379295037,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":127.87,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":366815218.1666667,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":4436483.5,"passRate":0.10476190476190476,"pass_rate":0.10476190476190476,"pass_rate_percent":10.476190476190476,"passes":21,"runs":201,"runtime":224862.70664021675,"source_flags":[],"source_harness":"openclaw","source_row_index":409,"source_status":"official_api_reported","split":"linux_only","splitTasks":105,"split_benchmark_ref":"agents-last-exam-v1-linux-only","split_tasks":105,"tasks":104,"tokens":null,"totalCostUsd":127.87,"totalDurationS":224862.70664021675,"totalInputTokens":366815218.1666667,"totalOutputTokens":4436483.5,"total_cost_usd":127.87,"total_input_tokens":366815218.1666667,"total_output_tokens":4436483.5,"total_runtime_s":224862.70664021675,"total_tokens":null,"track":"linux_only"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":10.476190476190476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-d01ab45c4ef494bc41ae","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2dde3e111e91f09888a164f2","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed/near-term","split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_track":"unlicensed/near-term","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.12686567164179105,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":1293,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[646];split=unlicensed/near-term;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":1293,"metadata":{"avgScore":0.36422336106940467,"avg_score":0.36422336106940467,"avg_score_percent":36.42233610694047,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":89.59,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":259732239.16666666,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":2654193.3333333335,"passRate":0.12686567164179105,"pass_rate":0.12686567164179105,"pass_rate_percent":12.686567164179104,"passes":17,"runs":131,"runtime":216764.1855777436,"source_flags":[],"source_harness":"openclaw","source_row_index":646,"source_status":"official_api_reported","split":"unlicensed/near-term","splitTasks":67,"split_benchmark_ref":"agents-last-exam-v1-unlicensed-near-term","split_tasks":67,"tasks":67,"tokens":null,"totalCostUsd":89.59,"totalDurationS":216764.1855777436,"totalInputTokens":259732239.16666666,"totalOutputTokens":2654193.3333333335,"total_cost_usd":89.59,"total_input_tokens":259732239.16666666,"total_output_tokens":2654193.3333333335,"total_runtime_s":216764.1855777436,"total_tokens":null,"track":"unlicensed/near-term"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":12.686567164179104,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-a1f5b568128d8217eb5e","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-2cef481fefca6c988c68e02a","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"licensed","split_benchmark_ref":"agents-last-exam-v1-licensed","split_track":"licensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":679,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[339];split=licensed;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":679,"metadata":{"avgScore":0.06469090909090909,"avg_score":0.06469090909090909,"avg_score_percent":6.469090909090909,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":40.36,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":120729225.5,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":575606.5,"passRate":0,"pass_rate":0,"pass_rate_percent":0.0,"passes":0,"runs":10,"runtime":46421.8629133185,"source_flags":[],"source_harness":"openclaw","source_row_index":339,"source_status":"official_api_reported","split":"licensed","splitTasks":11,"split_benchmark_ref":"agents-last-exam-v1-licensed","split_tasks":11,"tasks":7,"tokens":null,"totalCostUsd":40.36,"totalDurationS":46421.8629133185,"totalInputTokens":120729225.5,"totalOutputTokens":575606.5,"total_cost_usd":40.36,"total_input_tokens":120729225.5,"total_output_tokens":575606.5,"total_runtime_s":46421.8629133185,"total_tokens":null,"track":"licensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"agents-last-exam","benchmarkName":"Agents' Last Exam","benchmarkVersion":"ALE-v1","benchmarkVersionHint":null,"benchmarkVersionId":"agents-last-exam@v1","benchmarkVersionStatus":"source_reported","candidateId":"cand-agents-last-exam-993a377a0d7f16599f2c","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://agents-last-exam.org/api/demo/leaderboard","harness":"agents-last-exam","harnessId":"agents-last-exam","httpStatus":200,"id":"pub-0d7de1fc32c0eba8f706afd0","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate","modelName":"Qwen3.6 Plus","modelRef":"qwen3-6-plus","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","protocol":{"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost_source":"db","evaluator":"agents-last-exam","harness":"agents-last-exam","harness_variant":null,"source_harness":"openclaw","split":"unlicensed","split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_track":"unlicensed","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","harness_specific","missing_harness_variant","missing_source_date","split_specific","unregistered_source_harness"],"rank":null,"rawValue":0.09219858156028368,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f54c59f696d1a2a37a20027c9124197175e58821","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"agents-last-exam/candidates.jsonl","line":939,"payloadSha256":"e2ead78a77de0e7c2859ae79d758ed9de7c73cbc5bd4d935ac810905589bb1fb","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://agents-last-exam.org/api/demo/leaderboard","sourceId":"agents-last-exam","sourceLabel":"Agents' Last Exam · ALE-V1 leaderboard","sourceLocator":"rows[469];split=unlicensed;harness=openclaw;model=qwen3-6-plus;variant=unknown;metric=pass_rate","sourceMetricId":"pass_rate","sourcePageUrl":"https://agents-last-exam.org/leaderboard","sourceRow":{"artifact":"agents-last-exam/candidates.jsonl","line":939,"metadata":{"avgScore":0.26021994074929344,"avg_score":0.26021994074929344,"avg_score_percent":26.021994074929346,"benchmark_version":"ALE-v1","benchmark_version_id":"agents-last-exam@v1","cost":213.49,"costSource":"db","cost_source":"db","harness":"openclaw","harness_id":"agents-last-exam","harness_variant":null,"input_tokens":618181138.6666666,"metric_id":"pass_rate","metric_role":"primary","model":"qwen3-6-plus","output_tokens":6451781.833333333,"passRate":0.09219858156028368,"pass_rate":0.09219858156028368,"pass_rate_percent":9.219858156028367,"passes":25,"runs":262,"runtime":438548.9462786893,"source_flags":[],"source_harness":"openclaw","source_row_index":469,"source_status":"official_api_reported","split":"unlicensed","splitTasks":141,"split_benchmark_ref":"agents-last-exam-v1-unlicensed","split_tasks":141,"tasks":139,"tokens":null,"totalCostUsd":213.49,"totalDurationS":438548.9462786893,"totalInputTokens":618181138.6666666,"totalOutputTokens":6451781.833333333,"total_cost_usd":213.49,"total_input_tokens":618181138.6666666,"total_output_tokens":6451781.833333333,"total_runtime_s":438548.9462786893,"total_tokens":null,"track":"unlicensed"},"sourceStatus":"official_api_reported","source_flags":[]},"sourceSubjectType":"system","sourceUrl":"https://agents-last-exam.org/api/demo/leaderboard","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"percent","value":9.219858156028367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-13e64cd2a1473e82b7ba","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-46cca02714c1621ca8b1ad88","mappingCandidates":["deepseek/r1@0528"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek R1 (0528)","observedAt":"2025-06-06","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model deepseek/deepseek-reasoner","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":224},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":34.4,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9cb2e69ad41b848fa0fb00c0b33b016a69e724a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":57,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1485;dirname=2025-06-06-16-47-07--r1-diff","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":57,"metadata":{"aider_version":"0.84.1.dev","date":"2025-06-06","dirname":"2025-06-06-16-47-07--r1-diff","pass_num_1":77,"pass_num_2":160,"pass_rate_2":71.4},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":34.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-05-28","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7b8e878f593a31242713","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-425d6c2399ee46cc8bb6775f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-06-06","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"71.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0555afb91cd5d35830bcdd82e868160893dfd643","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":48,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=35;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":48,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":71.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-972ac78ef466fedf1aa6","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d219c0c0c28d1ff3735856d","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"70.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37b40a94e5a9556149905296e5dad677d3929ca4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":18,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=5;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":18,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":70.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-521cda904cd66445bb20","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-169fd93046d81f5a3c59baff","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp (Chat)","observedAt":"2025-10-03","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model deepseek/deepseek-chat","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":38.7,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"275165d1774aae5a133de55f5a92ecdedef2299b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":69,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1830;dirname=2025-10-03-09-21-36--deepseek-v3.2-chat","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":69,"metadata":{"aider_version":"0.86.2.dev","date":"2025-10-03","dirname":"2025-10-03-09-21-36--deepseek-v3.2-chat","pass_num_1":87,"pass_num_2":158,"pass_rate_2":70.2},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":38.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-fb59f927f1e371858165","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-a4eb189379d402d37b68fb7a","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp (Reasoner)","observedAt":"2025-10-03","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model deepseek/deepseek-reasoner","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":39.6,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"275165d1774aae5a133de55f5a92ecdedef2299b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":68,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1802;dirname=2025-10-03-09-45-34--deepseek-v3.2-reasoner","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":68,"metadata":{"aider_version":"0.86.2.dev","date":"2025-10-03","dirname":"2025-10-03-09-45-34--deepseek-v3.2-reasoner","pass_num_1":89,"pass_num_2":167,"pass_rate_2":74.2},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":39.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-681a0b64c39900e0fc41","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-789c924ce7c57319cc1ea595","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"74.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37b40a94e5a9556149905296e5dad677d3929ca4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":19,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=6;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":19,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":74.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-464fb15b24ece9454015","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3743508f7334322c858a2830","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"70.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26b97476924f4f7e6bcad6c162d1f44ef1fb0a18","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":36,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=23;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":36,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":70.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-020ac62e4aa807a8e6ca","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-747d737071598c375f75ae9b","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Aider Polyglot labels for the dated Claude Opus 4 release; thinking settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Aider Polyglot labels for the dated Claude Opus 4 release; thinking settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514 (32k thinking)","observedAt":"2025-05-25","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model claude-opus-4-20250514","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":37.3,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02039532a487c5e4645567d12039b826bdd6cd8a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":52,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1341;dirname=2025-05-25-20-40-51--opus4-diff-exuser","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":52,"metadata":{"aider_version":"0.83.3.dev","date":"2025-05-25","dirname":"2025-05-25-20-40-51--opus4-diff-exuser","pass_num_1":84,"pass_num_2":162,"pass_rate_2":72.0},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":37.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-77ef199fd08ea77c870a","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-41b74ea27a48b9872097532d","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Aider Polyglot labels for the dated Claude Opus 4 release; thinking settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Aider Polyglot labels for the dated Claude Opus 4 release; thinking settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514 (no think)","observedAt":"2025-05-25","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model claude-opus-4-20250514","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":32.9,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02039532a487c5e4645567d12039b826bdd6cd8a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":51,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1313;dirname=2025-05-25-19-57-20--opus4-diff-exuser","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":51,"metadata":{"aider_version":"0.83.3.dev","date":"2025-05-25","dirname":"2025-05-25-19-57-20--opus4-diff-exuser","pass_num_1":74,"pass_num_2":159,"pass_rate_2":70.7},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":32.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e9a9ae3ae36320c1143f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-da0d87078a1863e07665bf8e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_32K","observedAt":"2025-05-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"72.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26b97476924f4f7e6bcad6c162d1f44ef1fb0a18","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":37,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=24;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":37,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-12-01","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7a7e64d2823b377f7b9b","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a8e26eed3c55f8be3257e67a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"DeepSeek V3.2","modelRef":"deepseek-chat","observedAt":"2025-10-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"70.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4327f855623fca902466089e0cbcea6c28318c3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":87,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=78;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":87,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":70.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-e02f0753f47f3c1c63a1","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-af9d3ba6e98455ec5234fb01","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"GPT-5","modelRef":"gpt-5 (high)","observedAt":"2025-08-23","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model openai/gpt-5","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":52.0,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dc38540aa0953410428364be25f3f3e155eb509","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":65,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1715;dirname=2025-08-23-15-47-21--gpt-5-high","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":65,"metadata":{"aider_version":"0.86.2.dev","date":"2025-08-23","dirname":"2025-08-23-15-47-21--gpt-5-high","pass_num_1":117,"pass_num_2":198,"pass_rate_2":88.0},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":52.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-0a0b72d58c5798b52e3e","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-9581db8370c1cd126897f14e","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"GPT-5","modelRef":"gpt-5 (low)","observedAt":"2025-08-25","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model openai/gpt-5","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":43.1,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dc38540aa0953410428364be25f3f3e155eb509","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":67,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1773;dirname=2025-08-25-14-16-37--gpt-5-low","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":67,"metadata":{"aider_version":"0.86.2.dev","date":"2025-08-25","dirname":"2025-08-25-14-16-37--gpt-5-low","pass_num_1":97,"pass_num_2":183,"pass_rate_2":81.3},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":43.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-e27a0556033d8a36c28b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-8177701022e27455099ba475","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"GPT-5","modelRef":"gpt-5 (medium)","observedAt":"2025-08-25","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model openai/gpt-5","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":49.8,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dc38540aa0953410428364be25f3f3e155eb509","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":66,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1744;dirname=2025-08-25-13-23-27--gpt-5-medium","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":66,"metadata":{"aider_version":"0.86.2.dev","date":"2025-08-25","dirname":"2025-08-25-13-23-27--gpt-5-medium","pass_num_1":112,"pass_num_2":195,"pass_rate_2":86.7},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":49.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-63d0d4d66b9611904d51","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b245a84af2c8e2962cd551d7","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2aaaec672f4e61e01b27b571231cf57f14cb4466","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":22,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=9;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":22,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-aca79fc9b01db845dca4","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-92f31d48adc5718581501c1e","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"81.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2aaaec672f4e61e01b27b571231cf57f14cb4466","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":20,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=7;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":20,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":81.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-251d8ae524a5a4824f44","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-064247470914160460547c12","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Percent correct","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"86.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2aaaec672f4e61e01b27b571231cf57f14cb4466","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":21,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"aider_polyglot_external.csv:row=8;column=Percent correct","sourceMetricId":"Percent correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":21,"metadata":{"epoch_file":"aider_polyglot_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://aider.chat/docs/leaderboards/#polyglot-leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":86.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aider-polyglot","benchmarkName":"Aider Polyglot","benchmarkVersion":null,"benchmarkVersionHint":"aider-polyglot@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-aider-polyglot-16870dd69da6dff8f981","canonicalModelId":"xai/grok-4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","harness":"aider","harnessId":"aider","httpStatus":200,"id":"pub-1f67c794b8684edc8813b039","mappingCandidates":["xai/grok-4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"pass_rate_1","modelName":"Grok 4","modelRef":"grok-4 (high)","observedAt":"2025-07-11","parserVersion":"0.1.0","payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","protocol":{"command":"aider --model openrouter/x-ai/grok-4","edit_format":"diff","harness":"aider","subject_type":"system","test_cases":225},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":40.9,"retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41a849da14fcd0aa80ef0e7b2e8d27648ba6b15a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-aider-polyglot/candidates.jsonl","line":62,"payloadSha256":"85a50b25953512d18ba4bb0c23c0b8e626fcf9a5b52d287644b8a0b44b9535de","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","sourceId":"src-aider-polyglot","sourceLabel":"Aider · Polyglot coding leaderboard","sourceLocator":"yaml-row=1630;dirname=2025-07-11-19-37-40--xai-or-grok4-high","sourceMetricId":"pass_rate_1","sourcePageUrl":"https://aider.chat/docs/leaderboards/","sourceRow":{"artifact":"src-aider-polyglot/candidates.jsonl","line":62,"metadata":{"aider_version":"0.85.2.dev","date":"2025-07-11","dirname":"2025-07-11-19-37-40--xai-or-grok4-high","pass_num_1":92,"pass_num_2":179,"pass_rate_2":79.6},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/Aider-AI/aider/main/aider/website/_data/polyglot_leaderboard.yml","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":40.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-098a7fb7265de351e7e8","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=aime--aime_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2850b63e237adb7151156e20","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/MathArena--aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":19,"rawValue":90.83,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df45b586854fda63fe84c42d9e4026706476fa23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":19,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=19;row=18","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":19,"metadata":{"author":"Qwen","filename":".eval_results/MathArena--aime_2026.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":34,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=aime--aime_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-089dff9fb676b20750d1","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=aime--aime_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d410ddf9441c9cee04309c83","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/MathArena--aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":14,"rawValue":93.33,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"906035d0440a6f4ccc6164b760673917e5816270","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":14,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=14;row=13","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":14,"metadata":{"author":"Qwen","filename":".eval_results/MathArena--aime_2026.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":47,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=aime--aime_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-d95f8f5d06b66822ac76","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-388ec68b9de326d0244cf6bb","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":94.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8747478af9a7cc726e04104ea68b6e2044177657","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":13,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":13,"metadata":{"author":"Qwen","filename":".eval_results/aime_2026.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-029ace12cabc0ea71831","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-66bdd1d311b1b4ea5433c495","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":17,"rawValue":92.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f729676d71bebedd328f17459fd980de07540d95","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":17,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=17;row=16","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":17,"metadata":{"author":"Qwen","filename":".eval_results/aime_2026.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-efd002c870e3a6d4bd18","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=aime--aime_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-e603f3b5727cc0aaceb3eb93","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/MathArena--aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":6,"rawValue":95.83,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e8618b6229873aaac6b2204c8ef64a1368793d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":6,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=6;row=5","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":6,"metadata":{"author":"moonshotai","filename":".eval_results/MathArena--aime_2026.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":88,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=aime--aime_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-adbc308a3975d887995b","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-9a29ef28c0d2b42badd8191b","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":4,"rawValue":96.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f25f62532ea470cf3b5a3e753c01caab01dd77ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":4,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=4;row=3","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":4,"metadata":{"author":"moonshotai","filename":".eval_results/aime_2026.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-4e74f59a735254cdd10a","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=aime--aime_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-0a320e47ff1a3022741ca6c6","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/MathArena--aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":7,"rawValue":95.83,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84112de5dc0f5309c716881f33118821ae1d2a3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":7,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=7;row=6","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":7,"metadata":{"author":"zai-org","filename":".eval_results/MathArena--aime_2026.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=aime--aime_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"aime-2026","benchmarkName":"AIME 2026","benchmarkVersion":null,"benchmarkVersionHint":"aime-2026@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-aime-2026-5625313c47d2c6176127","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-11581d402e03644240b4d374","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","protocol":{"filename":".eval_results/MathArena--aime_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":11,"rawValue":95.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e49eddec3e6c7bd94514b76247ce7c8ba700d3c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-aime-2026/candidates.jsonl","line":11,"payloadSha256":"bd8f3e47321d266bfd4c817736e5f0d18cbcd5500ebef2c3b98cffdf31777d0c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-aime-2026","sourceLabel":"Hugging Face · AIME leaderboard API","sourceLocator":"rank=11;row=10","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","sourceRow":{"artifact":"hf-aime-2026/candidates.jsonl","line":11,"metadata":{"author":"zai-org","filename":".eval_results/MathArena--aime_2026.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/aime_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-75dc06ab2d3d57ddeefe","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-39dc9579246ff50a09053200","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Fable 5","modelRef":"Claude Fable 5 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":3,"rawValue":0.11615315314960728,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9cf318ac3c5e2f39c9f2becae579e6d871df2618","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2561,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=2","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2561,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1503802.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.11615315314960728,"score_ci_lower":0.09986286097855475,"score_ci_upper":0.13244344532065983,"session_count":32981.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.09986286097855475,"type":"confidence_interval","upper":0.13244344532065983}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.11615315314960728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-de84dab3745c6f551e58","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-6b6c74c43ad6f325eef8de24","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 4.6","modelRef":"Claude Opus 4.6","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":16,"rawValue":0.05775102458273259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f28b280ad102e725c070ed1d719fdd1fbb3fbac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2574,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=15","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2574,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1297746.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.05775102458273259,"score_ci_lower":0.04553400411908662,"score_ci_upper":0.06996804504637856,"session_count":36350.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.04553400411908662,"type":"confidence_interval","upper":0.06996804504637856}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.05775102458273259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-977f898b43c7acaacdd5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-a07e44f7b75ab7a1f5cb7ce3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 4.7","modelRef":"Claude Opus 4.7","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":10,"rawValue":0.07179330480418356,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e822d9d0102d92d635642e43abb80190808a1ad","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2568,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=9","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2568,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1466528.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.07179330480418356,"score_ci_lower":0.059037275241991365,"score_ci_upper":0.08454933436637575,"session_count":37099.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.059037275241991365,"type":"confidence_interval","upper":0.08454933436637575}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.07179330480418356,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bc57326e0098dd29eca0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-0ab61d1006c2cead37f70847","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 4.7","modelRef":"Claude Opus 4.7 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":9,"rawValue":0.0727341948352638,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e822d9d0102d92d635642e43abb80190808a1ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2567,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=8","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2567,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1381059.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.0727341948352638,"score_ci_lower":0.06011672147454766,"score_ci_upper":0.08535166819597995,"session_count":36563.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.06011672147454766,"type":"confidence_interval","upper":0.08535166819597995}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.0727341948352638,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-974553e184b4c1cacdc8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-36248fa8a206907425cf3b40","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 4.8","modelRef":"Claude Opus 4.8","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":23,"rawValue":0.0271912872310097,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2cd7ea357721b11f21161a176dcb2e6ed3370662","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2581,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=22","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2581,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1436791.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.0271912872310097,"score_ci_lower":0.0036375847271259455,"score_ci_upper":0.050744989734893456,"session_count":33678.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.0036375847271259455,"type":"confidence_interval","upper":0.050744989734893456}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.0271912872310097,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6227b0b3599f183fa4e0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-185586c3a17d46438ae69692","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 4.8","modelRef":"Claude Opus 4.8 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":5,"rawValue":0.09988008358694944,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2cd7ea357721b11f21161a176dcb2e6ed3370662","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2563,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=4","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2563,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1666284.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.09988008358694944,"score_ci_lower":0.08457303500796991,"score_ci_upper":0.11518713216592898,"session_count":35588.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.08457303500796991,"type":"confidence_interval","upper":0.11518713216592898}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.09988008358694944,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-80033260cfa39ed942ca","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-a23676fdea97d6886ea58468","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 5","modelRef":"Claude Opus 5 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":1,"rawValue":0.12733236681052468,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aad64f5a173294a86d18565b3339cc67289cd87f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2559,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=0","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2559,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":2365356.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.12733236681052468,"score_ci_lower":0.11104647158413665,"score_ci_upper":0.1436182620369127,"session_count":20070.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.11104647158413665,"type":"confidence_interval","upper":0.1436182620369127}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.12733236681052468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1a3ffb12443ece0fab47","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-775a2442eff50a6375fd7188","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Opus 5","modelRef":"Claude Opus 5 (Max)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":2,"rawValue":0.12413615325188568,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aad64f5a173294a86d18565b3339cc67289cd87f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2560,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=1","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2560,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1915828.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.12413615325188568,"score_ci_lower":0.10504709614339458,"score_ci_upper":0.14322521036037678,"session_count":15850.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.10504709614339458,"type":"confidence_interval","upper":0.14322521036037678}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.12413615325188568,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1fb904d658b7940eccd2","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-0a600fc00ac1f1e306ee523f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Sonnet 4.6","modelRef":"Claude Sonnet 4.6","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":24,"rawValue":0.022080938177991297,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"48efa46464194d68144bad44cabc9604e437d1a7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2582,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=23","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2582,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1270419.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.022080938177991297,"score_ci_lower":0.009586157217945554,"score_ci_upper":0.03457571913803704,"session_count":37002.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.009586157217945554,"type":"confidence_interval","upper":0.03457571913803704}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.022080938177991297,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-425205ff15fce8b85edb","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-38098b70b75d832bdd94f629","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Claude Sonnet 5","modelRef":"Claude Sonnet 5 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":8,"rawValue":0.08045201676788406,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7b080051f1fff0d255385de2414e13db5685ac5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2566,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=7","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2566,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":2538178.0,"organization":"anthropic","rating":null,"rating_lower":null,"rating_upper":null,"score":0.08045201676788406,"score_ci_lower":0.057719703546805945,"score_ci_upper":0.10318432998896218,"session_count":26108.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.057719703546805945,"type":"confidence_interval","upper":0.10318432998896218}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.08045201676788406,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f040c963a7ed498dfbd8","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-1beefbd1ac64e5f758bef238","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"DeepSeek V4 Pro","modelRef":"DeepSeek V4 Pro","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":30,"rawValue":-0.0018080733374592028,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a43c58ff6129ebfbaf65c4964a223f6f4ddc4a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2588,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=29","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2588,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":1445654.0,"organization":"deepseek","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.0018080733374592028,"score_ci_lower":-0.010909247600203925,"score_ci_upper":0.007293100925285519,"session_count":30472.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.010909247600203925,"type":"confidence_interval","upper":0.007293100925285519}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.0018080733374592028,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c81921ea954e6be55271","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-2095fa947d919162e67a1653","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Arena names the dated 0813 release and its High effort explicitly.","registry":"model_aliases.json"},"mappingNote":"Arena names the dated 0813 release and its High effort explicitly.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"DeepSeek V4 Pro (0813)","modelRef":"DeepSeek V4 Pro (High) (0813)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":11,"rawValue":0.06970690759579298,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6078ae567924440bfadad446b70a417ad7beacb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2569,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=10","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2569,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":1556406.0,"organization":"deepseek","rating":null,"rating_lower":null,"rating_upper":null,"score":0.06970690759579298,"score_ci_lower":0.05856703629921489,"score_ci_upper":0.08084677889237107,"session_count":18844.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.05856703629921489,"type":"confidence_interval","upper":0.08084677889237107}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.06970690759579298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-87ebe837dfbded438adf","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-aaa1cee900f2b65fc620748d","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Arena names the 2026-07-31 Flash release and High effort explicitly.","registry":"model_aliases.json"},"mappingNote":"Arena names the 2026-07-31 Flash release and High effort explicitly.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"DeepSeek V4 Flash","modelRef":"Deepseek V4 Flash (High) (20260731)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":21,"rawValue":0.0346379679398035,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ab323217d92aeeb20cdec8b09f90481fdd41c6d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2579,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=20","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2579,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":4443513.0,"organization":"deepseek","rating":null,"rating_lower":null,"rating_upper":null,"score":0.0346379679398035,"score_ci_lower":0.026845921505819947,"score_ci_upper":0.042430014373787045,"session_count":45881.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.026845921505819947,"type":"confidence_interval","upper":0.042430014373787045}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.0346379679398035,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0232b8d28869b482f662","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-1eb80a16affa6f47fd3f4c89","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GLM-5.1","modelRef":"GLM 5.1","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":32,"rawValue":-0.004246206481093473,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6c4b34b75f8251984653ebf9b8961ccf46759ae5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2590,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=31","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2590,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":3296127.0,"organization":"zai","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.004246206481093473,"score_ci_lower":-0.012352813418043063,"score_ci_upper":0.0038604004558561167,"session_count":71926.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.012352813418043063,"type":"confidence_interval","upper":0.0038604004558561167}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.004246206481093473,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b8d98ea5c8267a52db40","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-95dc89b59c846e61d36a6d3c","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GLM-5.2","modelRef":"GLM 5.2 (Max)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":15,"rawValue":0.05862951282046114,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0110f5e011f739ae3b1079789ae5c6afb5113060","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2573,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=14","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2573,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":2673775.0,"organization":"zai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.05862951282046114,"score_ci_lower":0.05034417290177853,"score_ci_upper":0.06691485273914374,"session_count":58668.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.05034417290177853,"type":"confidence_interval","upper":0.06691485273914374}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.05862951282046114,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5e17dcfc67f8c9ed0841","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-d38a265d501d169e799248ef","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GLM-5.3","modelRef":"GLM 5.3 (Max)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":20,"rawValue":0.03774351573659967,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59bbc3d85b7a1e39ef186979c5876d49a9056446","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2578,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=19","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2578,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":1338363.0,"organization":"zai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.03774351573659967,"score_ci_lower":0.026908126566415294,"score_ci_upper":0.04857890490678404,"session_count":22601.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.026908126566415294,"type":"confidence_interval","upper":0.04857890490678404}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.03774351573659967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-04485c6b3ab07f28662f","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-aa680db93269d7e5d51e4500","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.4","modelRef":"GPT 5.4 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":18,"rawValue":0.045811011871245894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d608140c2cf13b74eddcf2287d1af483ec343f19","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2576,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=17","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2576,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":3162187.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.045811011871245894,"score_ci_lower":0.03749814853506049,"score_ci_upper":0.054123875207431296,"session_count":74927.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.03749814853506049,"type":"confidence_interval","upper":0.054123875207431296}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.045811011871245894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3de5f5eccac402fe5132","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-1d29226e7ab56b9b62bc42bf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.5","modelRef":"GPT 5.5","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":17,"rawValue":0.056818461920492905,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82658ec269a12d26edb91a5e10a659ee879cd592","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2575,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=16","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2575,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1804269.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.056818461920492905,"score_ci_lower":0.048653784058741746,"score_ci_upper":0.06498313978224406,"session_count":75680.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.048653784058741746,"type":"confidence_interval","upper":0.06498313978224406}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.056818461920492905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0afdaadf0fba5bbc8551","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-233911bc4dfe5e603d3d94dc","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.5","modelRef":"GPT 5.5 (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":12,"rawValue":0.06758524664177534,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82658ec269a12d26edb91a5e10a659ee879cd592","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2570,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=11","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2570,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":2264598.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.06758524664177534,"score_ci_lower":0.058636868013626456,"score_ci_upper":0.07653362526992423,"session_count":73820.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.058636868013626456,"type":"confidence_interval","upper":0.07653362526992423}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.06758524664177534,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-024d359b8503ea78f477","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-73325ffdf060eabbec264654","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.5","modelRef":"GPT 5.5 (xHigh)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":7,"rawValue":0.0855784909898381,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82658ec269a12d26edb91a5e10a659ee879cd592","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2565,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=6","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2565,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":2099912.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.0855784909898381,"score_ci_lower":0.0759202356795503,"score_ci_upper":0.0952367463001259,"session_count":48549.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.0759202356795503,"type":"confidence_interval","upper":0.0952367463001259}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.0855784909898381,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0540370e588e4ea021fe","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-41ceb54beb7d732578f2af3f","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.6 Luna","modelRef":"GPT 5.6 Luna (xHigh)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":19,"rawValue":0.03829833242271487,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3532af6b61478c8a38bd046f31547e4a41603cd9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2577,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=18","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2577,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":777846.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.03829833242271487,"score_ci_lower":0.022883249916946364,"score_ci_upper":0.053713414928483375,"session_count":11919.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.022883249916946364,"type":"confidence_interval","upper":0.053713414928483375}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.03829833242271487,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-050c4aab63ae3dd56781","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-28961e9a3695953a508666ca","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.6 Sol","modelRef":"GPT 5.6 Sol (xHigh)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":4,"rawValue":0.10308907749743668,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f1a5e7a7b75ec483be09ab6191d630067e4311e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2562,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=3","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2562,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":2642492.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.10308907749743668,"score_ci_lower":0.08856131140022662,"score_ci_upper":0.11761684359464673,"session_count":26960.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.08856131140022662,"type":"confidence_interval","upper":0.11761684359464673}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.10308907749743668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-89aaa8992206cbc6ad9d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-f63e920b5a207f21d088a53e","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"GPT-5.6 Terra","modelRef":"GPT 5.6 Terra (xHigh)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":22,"rawValue":0.03298223715667077,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a528abc214296cfdc7d189d2f3439a152c29803e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2580,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=21","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2580,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":834478.0,"organization":"openai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.03298223715667077,"score_ci_lower":0.02125859980578569,"score_ci_upper":0.04470587450755585,"session_count":15268.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.02125859980578569,"type":"confidence_interval","upper":0.04470587450755585}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.03298223715667077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-65913abed93e03ec12a8","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-37b44338a1cbefddbd1dedd4","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Gemini 3.1 Pro","modelRef":"Gemini 3.1 Pro Preview","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":36,"rawValue":-0.024091974587494024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1920e2b234f6c0cff8cdd25bdde37e405d39f63b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2594,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=35","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2594,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":2626762.0,"organization":"google","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.024091974587494024,"score_ci_lower":-0.031503677670357555,"score_ci_upper":-0.016680271504630494,"session_count":81874.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.031503677670357555,"type":"confidence_interval","upper":-0.016680271504630494}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.024091974587494024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-61e2d88dfdcabcfef323","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-297fe35d70467c0afff797cb","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Gemini 3.5 Flash","modelRef":"Gemini 3.5 Flash (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":35,"rawValue":-0.02221800402823942,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d308ff399ef35d62be1c060394a791bb9678f98","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2593,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=34","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2593,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":4114882.0,"organization":"google","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.02221800402823942,"score_ci_lower":-0.02901112385552642,"score_ci_upper":-0.01542488420095242,"session_count":94392.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.02901112385552642,"type":"confidence_interval","upper":-0.01542488420095242}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.02221800402823942,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-306be88458b5c53552e0","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-24bedef68563fb85d35e1f9d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Gemini 3.5 Flash","modelRef":"Gemini 3.5 Flash (Medium)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":40,"rawValue":-0.03484727666575396,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d308ff399ef35d62be1c060394a791bb9678f98","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2598,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=39","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2598,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":562752.0,"organization":"google","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.03484727666575396,"score_ci_lower":-0.04886343797012255,"score_ci_upper":-0.020831115361385374,"session_count":13738.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.04886343797012255,"type":"confidence_interval","upper":-0.020831115361385374}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.03484727666575396,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f4fe259d9a6314c46970","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-174cd2be9c579ed18b507ab5","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Gemini 3.5 Flash-Lite","modelRef":"Gemini 3.5 Flash Lite","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":49,"rawValue":-0.10872946128446538,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d1b3bd1478a0fe75465aef1dd7975e110ec5671","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2607,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=48","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2607,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":577510.0,"organization":"google","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.10872946128446538,"score_ci_lower":-0.12205644743317776,"score_ci_upper":-0.095402475135753,"session_count":20878.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.12205644743317776,"type":"confidence_interval","upper":-0.095402475135753}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.10872946128446538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-50dcac59cf4aaaf51edb","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-d027fab2c1ae71dd4dc81a1e","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Gemini 3.6 Flash","modelRef":"Gemini 3.6 Flash (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":39,"rawValue":-0.0321966465690072,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a1729fb1df680cb1777d4b5cbcc8e0d9ead98ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2597,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=38","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2597,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":775103.0,"organization":"google","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.0321966465690072,"score_ci_lower":-0.04437750522831556,"score_ci_upper":-0.020015787909698837,"session_count":14960.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.04437750522831556,"type":"confidence_interval","upper":-0.020015787909698837}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.0321966465690072,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3e0e48c1928736c81c3e","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-3f6f3479d5a9d7524ca78092","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Gemini 3.7 Flash","modelRef":"Gemini 3.7 Flash (High)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":27,"rawValue":0.015668534602611646,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d77d947c93c27603af679f3c97338ef0d147914","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2585,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=26","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2585,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1726049.0,"organization":"google","rating":null,"rating_lower":null,"rating_upper":null,"score":0.015668534602611646,"score_ci_lower":0.006913731469758223,"score_ci_upper":0.024423337735465067,"session_count":24091.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.006913731469758223,"type":"confidence_interval","upper":0.024423337735465067}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.015668534602611646,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4f3a951ac2100ad2b4bc","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-2adf258d87b8ea738b3bdba7","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Grok 4.5","modelRef":"Grok 4.5","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":14,"rawValue":0.06384994584206044,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9bb76fe9286df89d300ac86c6237fcacbef5a37","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2572,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=13","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2572,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1527238.0,"organization":"xai","rating":null,"rating_lower":null,"rating_upper":null,"score":0.06384994584206044,"score_ci_lower":0.05202747950909302,"score_ci_upper":0.07567241217502786,"session_count":32149.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.05202747950909302,"type":"confidence_interval","upper":0.07567241217502786}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.06384994584206044,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-86138d268817664de509","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-fffa0dfb5f17c783f8eebc75","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Kimi K2.6","modelRef":"Kimi K2.6","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":28,"rawValue":0.010858661242663967,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf3edd38eab65576d4d762ac09aaa767f0e5ad52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2586,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=27","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2586,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":327644.0,"organization":"moonshot","rating":null,"rating_lower":null,"rating_upper":null,"score":0.010858661242663967,"score_ci_lower":-0.012866063573805674,"score_ci_upper":0.03458338605913361,"session_count":11272.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.012866063573805674,"type":"confidence_interval","upper":0.03458338605913361}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.010858661242663967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5ccc2751bbe265730d56","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-7961fe5c7ea5159eb206c8ed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Kimi K2.7 Code","modelRef":"Kimi K2.7 Code","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":25,"rawValue":0.020284108795353084,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f375de537c3d94cc88adfbce665a8dffcd70e83","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2583,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=24","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2583,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":409470.0,"organization":"moonshot","rating":null,"rating_lower":null,"rating_upper":null,"score":0.020284108795353084,"score_ci_lower":-0.001485813350425358,"score_ci_upper":0.042054030941131526,"session_count":11137.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.001485813350425358,"type":"confidence_interval","upper":0.042054030941131526}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.020284108795353084,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4032da2f86af25180666","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-9ef4b938deda4522609c9f3c","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Kimi K3","modelRef":"Kimi K3 (Max)","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":6,"rawValue":0.09531968013602259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3083b497dbdcfdfb85d977f9c809757a40c009ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2564,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=5","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2564,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":7284695.0,"organization":"moonshot","rating":null,"rating_lower":null,"rating_upper":null,"score":0.09531968013602259,"score_ci_lower":0.08917553427833186,"score_ci_upper":0.10146382599371331,"session_count":89758.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.08917553427833186,"type":"confidence_interval","upper":0.10146382599371331}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.09531968013602259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0f892c7b8126d338d265","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-1f43b0d04e214fe8bed28f2b","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"MiMo-V2.5-Pro","modelRef":"Mimo V2.5 Pro","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":37,"rawValue":-0.02787033746291722,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"166ab65feea051ba8ad1f15f5aeb316a734f1fec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2595,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=36","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2595,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":1458279.0,"organization":"xiaomi","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.02787033746291722,"score_ci_lower":-0.036953399885430865,"score_ci_upper":-0.018787275040403575,"session_count":34082.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.036953399885430865,"type":"confidence_interval","upper":-0.018787275040403575}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.02787033746291722,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3534132551a1286ddad5","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-1a2ee37e073afa42c53b89c6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"MiniMax M2.7","modelRef":"Minimax M2.7","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":47,"rawValue":-0.1015768677145034,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29eba06dc804c198d958f5dcec80b2b8f6f699bb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2605,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=46","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2605,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":737494.0,"organization":"minimax","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.1015768677145034,"score_ci_lower":-0.11330946932093995,"score_ci_upper":-0.08984426610806685,"session_count":33743.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.11330946932093995,"type":"confidence_interval","upper":-0.08984426610806685}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.1015768677145034,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3c8f47ef2c5e0cff3304","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-d6685c16766adf319d91d735","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"MiniMax M3","modelRef":"Minimax M3","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":38,"rawValue":-0.031007512966943346,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4a49367e4f78fe8268b2ff04482fdd5445538c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2596,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=37","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2596,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":2269108.0,"organization":"minimax","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.031007512966943346,"score_ci_lower":-0.039668950285115724,"score_ci_upper":-0.022346075648770967,"session_count":33558.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.039668950285115724,"type":"confidence_interval","upper":-0.022346075648770967}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.031007512966943346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3345d3219e31d9775d2c","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-16122c25fd287e99a4e1a9aa","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Qwen3.7 Max","modelRef":"Qwen3.7 Max","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","curated_model_alias","inferred_benchmark_version_hint"],"rank":33,"rawValue":-0.010807488639950926,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85289a6a2f00f79e62b3bc854d0d2aa96557d0eb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2591,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=32","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2591,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1422232.0,"organization":"alibaba","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.010807488639950926,"score_ci_lower":-0.01975230888372835,"score_ci_upper":-0.0018626683961735016,"session_count":33252.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.01975230888372835,"type":"confidence_interval","upper":-0.0018626683961735016}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.010807488639950926,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0d8ee215cc6d75f44dc8","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-f8554e026ed13e81144f91f3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Qwen3.7 Plus","modelRef":"Qwen3.7 Plus","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":34,"rawValue":-0.020675286771546914,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"331a3787794868755587395e110048e378a6f246","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2592,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=33","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2592,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":743416.0,"organization":"alibaba","rating":null,"rating_lower":null,"rating_upper":null,"score":-0.020675286771546914,"score_ci_lower":-0.034003559547728775,"score_ci_upper":-0.0073470139953650555,"session_count":17443.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":-0.034003559547728775,"type":"confidence_interval","upper":-0.0073470139953650555}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":-0.020675286771546914,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-agent","benchmarkName":"Agent Arena","benchmarkVersion":null,"benchmarkVersionHint":"arena-agent@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f2a7746912b97d159303","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-agent","harnessId":"arena-agent","httpStatus":200,"id":"pub-95b34f738cacc70ed76fe387","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ips","modelName":"Qwen3.8 Max","modelRef":"Qwen3.8 Max","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"agent","category":"overall","harness":"arena-agent","rating_method":"ips","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_ips","inferred_benchmark_version_hint"],"rank":13,"rawValue":0.06459728013665357,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57de0bfff1cefdc308d39a0879888ba0421477f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2571,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=agent;split=latest;row_idx=12","sourceMetricId":"ips","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2571,"metadata":{"category":"overall","config":"agent","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":1023424.0,"organization":"alibaba","rating":null,"rating_lower":null,"rating_upper":null,"score":0.06459728013665357,"score_ci_lower":0.052265614102247476,"score_ci_upper":0.07692894617105966,"session_count":14712.0,"source_status":"publisher_dataset","split":"latest","variance":null,"vote_count":null},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":0.052265614102247476,"type":"confidence_interval","upper":0.07692894617105966}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:kind=agent-eval"],"subjectType":"system","unit":"fraction","value":0.06459728013665357,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2e29696de2b7816ec881","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4df5941c9e64215e671e3c62","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1504.2491537844257,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"742998687132e799eb85c4250bafd15cab10cb25","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2524,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=3","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2524,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1504.2491537844257,"rating_lower":1495.079648738949,"rating_upper":1513.4186588299021,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.887472104627797,"vote_count":5792},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1495.079648738949,"type":"confidence_interval","upper":1513.4186588299021}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1504.2491537844257,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fafd5254832437a264ce","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-69cba88238aba653266e586b","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1423.0272414121418,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d31d2429ad09e5bf1afe192d3ba0d1f4cfc6ce15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2550,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=29","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2550,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1423.0272414121418,"rating_lower":1416.5957079036052,"rating_upper":1429.4587749206785,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.767946554208846,"vote_count":30937},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1416.5957079036052,"type":"confidence_interval","upper":1429.4587749206785}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1423.0272414121418,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3bc1fd6d789019013d3e","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c5aada10249c7222dec885ce","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1461.6531335915963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f185b93326ff0d2b5690082a6c1982dc04f7ce2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2539,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=18","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2539,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1461.6531335915963,"rating_lower":1451.2244273921856,"rating_upper":1472.081839791007,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.31161755730449,"vote_count":7965},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1451.2244273921856,"type":"confidence_interval","upper":1472.081839791007}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1461.6531335915963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-22a311e39154dd54c0dc","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-997a857cf88bb99c7af5e904","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1509.9436313705492,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f74f7f998adf0b69d7a9b462e3e10f0e587320b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2522,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=1","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2522,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1509.9436313705492,"rating_lower":1503.540386895363,"rating_upper":1516.3468758457352,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.673429476355041,"vote_count":37271},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1503.540386895363,"type":"confidence_interval","upper":1516.3468758457352}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1509.9436313705492,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-289df0522f8845301450","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-663aca0565d1905fe272e17d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1505.9121170864287,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f74f7f998adf0b69d7a9b462e3e10f0e587320b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2523,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=2","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2523,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1505.9121170864287,"rating_lower":1498.8564635694056,"rating_upper":1512.9677706034518,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.959203489075481,"vote_count":24973},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1498.8564635694056,"type":"confidence_interval","upper":1512.9677706034518}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1505.9121170864287,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c6d750cdb3c9a04ddaa9","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ac3c6df57a50e69c0a42154d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1498.1450675830365,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da646730912eb934f27916d94809494cb98c53c4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2525,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=4","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2525,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1498.1450675830365,"rating_lower":1491.0689343067083,"rating_upper":1505.2212008593651,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.034543511080322,"vote_count":18990},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1491.0689343067083,"type":"confidence_interval","upper":1505.2212008593651}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1498.1450675830365,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dc2e6bf98e2a986ad191","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aadf5e721d95227f8da1de7c","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-thinking","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1497.0210273683192,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da646730912eb934f27916d94809494cb98c53c4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2526,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=5","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2526,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1497.0210273683192,"rating_lower":1489.908182646685,"rating_upper":1504.1338720899535,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.170142488977906,"vote_count":18793},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1489.908182646685,"type":"confidence_interval","upper":1504.1338720899535}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1497.0210273683192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a2321a2f26febf1021cd","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fa0293f3d521f4efc7b6f770","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1468.9024272352567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c63c8846fb3482b1370993abde186bc9ed061fbb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2536,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=15","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2536,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1468.9024272352567,"rating_lower":1460.5800927551188,"rating_upper":1477.2247617153944,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.0299345722981,"vote_count":8193},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.5800927551188,"type":"confidence_interval","upper":1477.2247617153944}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1468.9024272352567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf61b16cd9aba0ef6cda","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-492cf215db25aaa017945e21","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-thinking","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1474.5889982256267,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c63c8846fb3482b1370993abde186bc9ed061fbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2532,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=11","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2532,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1474.5889982256267,"rating_lower":1466.4523131662731,"rating_upper":1482.7256832849807,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.234505651461575,"vote_count":8388},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1466.4523131662731,"type":"confidence_interval","upper":1482.7256832849807}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1474.5889982256267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-af89ff46efde7c43dbce","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c932546fbf518850605ce7e5","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1520.1085960110886,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22eda0748a3b6dbb74458c048e738989605b0975","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2521,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=0","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2521,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1520.1085960110886,"rating_lower":1505.483506741573,"rating_upper":1534.7336852806047,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":55.68021059839379,"vote_count":1663},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1505.483506741573,"type":"confidence_interval","upper":1534.7336852806047}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1520.1085960110886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a4050b243962a87edab9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0b53fafca7877cc9bdb559c7","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1446.1948727501483,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c6862664fc659f013a0c0bc21ecef94d7f0c35a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2542,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=21","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2542,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1446.1948727501483,"rating_lower":1439.6000407270108,"rating_upper":1452.7897047732858,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.32169091052311,"vote_count":28634},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1439.6000407270108,"type":"confidence_interval","upper":1452.7897047732858}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1446.1948727501483,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2445ee0387d216fba58a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cd8d406f9107c4f27bfb49cb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1482.7958418713733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea2fae94daa13de63bee22d618dc401126754d99","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2528,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=7","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2528,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1482.7958418713733,"rating_lower":1476.4979983740732,"rating_upper":1489.093685368674,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.32494022917075,"vote_count":54427},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1476.4979983740732,"type":"confidence_interval","upper":1489.093685368674}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1482.7958418713733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2005e25858c7233edc86","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-50b582c40453395af34f0acf","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1470.2233955126517,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9a434d4b1993ec26c63d3e3f0468af3d3a3c28b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2534,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=13","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2534,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1470.2233955126517,"rating_lower":1459.8485910113407,"rating_upper":1480.5982000139625,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.0197116419881,"vote_count":3601},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1459.8485910113407,"type":"confidence_interval","upper":1480.5982000139625}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1470.2233955126517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-08186fb03b97478d76c8","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-230ef73a7dc629b18a89ba84","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1420.8717297728372,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c2b9df4d392039f6cc35826bb223b141a1321aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2551,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=30","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2551,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1420.8717297728372,"rating_lower":1414.583287180114,"rating_upper":1427.1601723655604,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.294138786271846,"vote_count":24963},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1414.583287180114,"type":"confidence_interval","upper":1427.1601723655604}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1420.8717297728372,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e3b9098de9c92829d11b","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-221c089ad7f939a9b220faa2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1445.3243879762572,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"239c2163a1df070018eb74ddb8e61021583f95d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2543,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=22","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2543,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1445.3243879762572,"rating_lower":1439.355384494359,"rating_upper":1451.2933914581554,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.274862553511207,"vote_count":45162},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1439.355384494359,"type":"confidence_interval","upper":1451.2933914581554}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1445.3243879762572,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c14e825facb1b80ccbcc","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7a8d7bec3cb2ad8dd74239b9","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1464.646724197627,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a2c54760ec540304f2ecf659823a4bf00d5308a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2537,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=16","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2537,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1464.646724197627,"rating_lower":1454.409538373915,"rating_upper":1474.8839100213395,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.28129559131399,"vote_count":3689},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1454.409538373915,"type":"confidence_interval","upper":1474.8839100213395}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.646724197627,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3c935a4480a0535a1ebb","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d036ec92bc673ed59864ab9b","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1417.1818912745823,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7ddc1ff1e6fdd622dcebfb2fd082e814c092bb5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2552,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=31","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2552,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1417.1818912745823,"rating_lower":1407.6431986831753,"rating_upper":1426.7205838659893,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.68544362969913,"vote_count":4894},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1407.6431986831753,"type":"confidence_interval","upper":1426.7205838659893}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1417.1818912745823,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0c6ab59047db2e398cd2","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ce6998704b46d0b27233b594","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1400.8394066066326,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"65b89bc5fb204fd6eef34acaa3799dd71222593b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2557,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=36","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2557,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1400.8394066066326,"rating_lower":1391.4014437462545,"rating_upper":1410.2773694670109,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.187842721110616,"vote_count":8220},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1391.4014437462545,"type":"confidence_interval","upper":1410.2773694670109}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1400.8394066066326,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-98c27d05779b02fb909b","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-94c6a8bf8bb5500e57e7de14","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1400.0724185714537,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"485f01d53ee452b7d39cb0dc47aa141113a00daa","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2558,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=37","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2558,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1400.0724185714537,"rating_lower":1394.127373376354,"rating_upper":1406.0174637665534,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.200557397981486,"vote_count":28068},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.127373376354,"type":"confidence_interval","upper":1406.0174637665534}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1400.0724185714537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2ef963769d59d388aa1a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d9b6d59cf14ec666793c9db7","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1405.0199151746874,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"485f01d53ee452b7d39cb0dc47aa141113a00daa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2555,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=34","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2555,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1405.0199151746874,"rating_lower":1395.4893851237312,"rating_upper":1414.5504452256441,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.64492430918998,"vote_count":7073},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1395.4893851237312,"type":"confidence_interval","upper":1414.5504452256441}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1405.0199151746874,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a829268be42392d9bd0d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-62aa69a9861b052029162d41","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1470.1387900819968,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad42da1136d4c9518eb614257e601a64bd03bb90","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2535,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=14","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2535,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1470.1387900819968,"rating_lower":1463.5435719115558,"rating_upper":1476.7340082524383,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.32301678763233,"vote_count":29809},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1463.5435719115558,"type":"confidence_interval","upper":1476.7340082524383}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1470.1387900819968,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5c7aba2b665722be1045","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-83a12c0d67d9272abf955576","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1480.2983593519762,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f63d8be6fb4c2c08d39a2b936e67d5c73abaceef","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2529,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=8","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2529,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1480.2983593519762,"rating_lower":1473.4648838292946,"rating_upper":1487.1318348746581,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.155899594063989,"vote_count":17286},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1473.4648838292946,"type":"confidence_interval","upper":1487.1318348746581}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1480.2983593519762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3255f8f6b6b24b711744","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a8d48587c4d6b3f0309265f3","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1484.7040517486435,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f63d8be6fb4c2c08d39a2b936e67d5c73abaceef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2527,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=6","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2527,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1484.7040517486435,"rating_lower":1477.6934924651682,"rating_upper":1491.7146110321191,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.794082602775932,"vote_count":16816},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1477.6934924651682,"type":"confidence_interval","upper":1491.7146110321191}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1484.7040517486435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e2e830b6d6ba90097080","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b4387ff2af571950f29f48a2","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1401.596483346334,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f63d8be6fb4c2c08d39a2b936e67d5c73abaceef","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2556,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=35","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2556,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1401.596483346334,"rating_lower":1393.6035165977628,"rating_upper":1409.589450094905,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.631056175742383,"vote_count":8442},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1393.6035165977628,"type":"confidence_interval","upper":1409.589450094905}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1401.596483346334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1c8130d02013630c0f60","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-325f4b41894a283d531b6fc5","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1462.304000455754,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4fffbebcfe21a81f068bfd32584ed2f6af6c304","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2538,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=17","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2538,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1462.304000455754,"rating_lower":1449.0213338320582,"rating_upper":1475.5866670794499,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":45.927664689730705,"vote_count":1888},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1449.0213338320582,"type":"confidence_interval","upper":1475.5866670794499}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1462.304000455754,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2168a49d9a8088870557","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-307c16e3da4c7ade01749734","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1478.6820927474027,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ccc800f837db326df840ba09f3f42a8ce68fa20","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2531,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=10","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2531,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1478.6820927474027,"rating_lower":1461.9414492175033,"rating_upper":1495.4227362773022,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":72.95383313376813,"vote_count":1152},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1461.9414492175033,"type":"confidence_interval","upper":1495.4227362773022}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1478.6820927474027,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cfc38a57164c97c0c62b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-66deaf875306c705d625770a","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1479.2063866025728,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"207e43b3ea6c40f9c0cfd601a11daa042e8d4355","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2530,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=9","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2530,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1479.2063866025728,"rating_lower":1466.1933142356384,"rating_upper":1492.2194589695073,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.08222509501567,"vote_count":1969},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1466.1933142356384,"type":"confidence_interval","upper":1492.2194589695073}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1479.2063866025728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3343bc3e997a89f39279","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e264b6734df427514f3bf02a","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1453.732648558952,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ee1dafd0b00d6eed35bd5507a3b390fe51d76a7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2540,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=19","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2540,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1453.732648558952,"rating_lower":1441.9374433392036,"rating_upper":1465.5278537787003,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.217195776379974,"vote_count":2435},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1441.9374433392036,"type":"confidence_interval","upper":1465.5278537787003}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1453.732648558952,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f2782d6ca28c474484ee","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-67af861372a31bd3aac819ca","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1429.4883922629656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90b5e419dff8a72367c1437704b5a3bb97fa8065","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2548,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=27","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2548,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1429.4883922629656,"rating_lower":1422.6388739472645,"rating_upper":1436.3379105786667,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.21304284309522,"vote_count":19891},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1422.6388739472645,"type":"confidence_interval","upper":1436.3379105786667}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1429.4883922629656,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4d214ee3f8d7cd344623","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8d21a454c47e939ad50e5e86","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1450.9009906110664,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8672a71d3b4b544e564630981336b7292334f568","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2541,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=20","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2541,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1450.9009906110664,"rating_lower":1443.2187274504354,"rating_upper":1458.583253771697,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.363217471252542,"vote_count":11181},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1443.2187274504354,"type":"confidence_interval","upper":1458.583253771697}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1450.9009906110664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2cb4da901e9bae75e4c7","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e3cbd881eb2946fd3c33783f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1434.8163626138064,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28da6f125bdcd23126e79c696c7f8ecf0ff54297","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2546,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=25","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2546,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1434.8163626138064,"rating_lower":1426.4529723892242,"rating_upper":1443.1797528383886,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.208263920943956,"vote_count":6263},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1426.4529723892242,"type":"confidence_interval","upper":1443.1797528383886}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1434.8163626138064,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-document","benchmarkName":"Arena Document","benchmarkVersion":null,"benchmarkVersionHint":"arena-document@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b588f2e1422b33da4642","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c261d4bfae667a61037f8d65","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-07-30","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"document","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1439.5689281931259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4995745934c53400c3ec6d81a26582da1251fee6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2545,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=document;split=latest;row_idx=24","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2545,"metadata":{"category":"overall","config":"document","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1439.5689281931259,"rating_lower":1428.959365661719,"rating_upper":1450.1784907245328,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.302101717569286,"vote_count":3088},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1428.959365661719,"type":"confidence_interval","upper":1450.1784907245328}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1439.5689281931259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-474cbcce044fe028669e","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-08dcdd8ac4eb7c6249bdd972","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1230.2225243963933,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9aafebfd29db787ab791d4f89a890d5cfd5e98a1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2491,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=4","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2491,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1230.2225243963933,"rating_lower":1222.3089533463658,"rating_upper":1238.1360954464205,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.302298081778126,"vote_count":41795},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1222.3089533463658,"type":"confidence_interval","upper":1238.1360954464205}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1230.2225243963933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-705efdf39ac6d491ff2a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-44215c724c2d3e739e854421","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1179.6445104274662,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"569e15c6faecacd4c23e90ff2afeb5e276bbdd94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2505,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=18","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2505,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1179.6445104274662,"rating_lower":1174.028720507639,"rating_upper":1185.2603003472937,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.209666664586468,"vote_count":61573},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1174.028720507639,"type":"confidence_interval","upper":1185.2603003472937}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1179.6445104274662,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-132d2cfce3d06db8d2ee","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8880b1a192080bc08d70e837","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1253.415899373882,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1b900e18dd577162af6bb23e837518bc4833111c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2488,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=1","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2488,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1253.415899373882,"rating_lower":1248.4459416557172,"rating_upper":1258.3858570920468,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.429973838918245,"vote_count":134699},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1248.4459416557172,"type":"confidence_interval","upper":1258.3858570920468}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1253.415899373882,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-40d61efa37096bdc4c22","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-01521887ac72987f26e403ef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1233.2545271621282,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7666197eb9372889580c7120e632e5c1173d4d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2490,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=3","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2490,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1233.2545271621282,"rating_lower":1227.873748107417,"rating_upper":1238.6353062168391,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.536924014295596,"vote_count":91394},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1227.873748107417,"type":"confidence_interval","upper":1238.6353062168391}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1233.2545271621282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-55e11be28ba1475da4c1","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-745521a5bf4342fbbcfd5fec","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1204.3009106749619,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca15c5c04e67aa2858429bb2733038500dfb65da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2498,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=11","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2498,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1204.3009106749619,"rating_lower":1197.8911268368415,"rating_upper":1210.7106945130822,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.695241253166913,"vote_count":70998},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1197.8911268368415,"type":"confidence_interval","upper":1210.7106945130822}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1204.3009106749619,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b0c428acd33e444f5330","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-78ef03ff10eddcf9033a34b7","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena search endpoint presentation of the catalogued Claude Opus 4 release.","registry":"model_aliases.json"},"mappingNote":"Arena search endpoint presentation of the catalogued Claude Opus 4 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1126.4656766180015,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74fbd63b031ebeba174b702016a38adc8031560f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2518,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=31","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2518,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1126.4656766180015,"rating_lower":1121.1434400917383,"rating_upper":1131.7879131442646,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.3738136899698175,"vote_count":31037},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1121.1434400917383,"type":"confidence_interval","upper":1131.7879131442646}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1126.4656766180015,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ca59f6d4d42b2a9b156b","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0c82d64fe143a67d13def394","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1158.4558124801003,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b54d08ffdc3801a54cfcfcd1aa7c62bb638c729","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2510,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=23","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2510,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1158.4558124801003,"rating_lower":1153.4144752885884,"rating_upper":1163.4971496716125,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.6159971679533625,"vote_count":127378},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1153.4144752885884,"type":"confidence_interval","upper":1163.4971496716125}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1158.4558124801003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-51485b212d652d56708c","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-152575b3fc26f7a1bcc301f9","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1221.231466167688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36b44efcb2ebea3f71a18d0f1e24f2cb51850b85","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2493,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=6","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2493,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1221.231466167688,"rating_lower":1216.394720706573,"rating_upper":1226.068211628803,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.089901713794683,"vote_count":134905},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1216.394720706573,"type":"confidence_interval","upper":1226.068211628803}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1221.231466167688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f5ebfc2719cca0d8aa32","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-627a682e57754a3ab9184e1a","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1194.0719510316353,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"818e08ff86fada5eee32dc4e7a920b79a478b58c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2503,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=16","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2503,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1194.0719510316353,"rating_lower":1187.1804054070499,"rating_upper":1200.9634966562207,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.36337634023139,"vote_count":40230},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1187.1804054070499,"type":"confidence_interval","upper":1200.9634966562207}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1194.0719510316353,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0fcbdde965b1ebb85495","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-16571658850c79f16d3ab777","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena search endpoint presentation of GPT-5; search/tool context remains in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena search endpoint presentation of GPT-5; search/tool context remains in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1132.767611287441,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"493ea009fa6fe72233c78ecce88b42fa9387c0c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2516,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=29","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2516,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1132.767611287441,"rating_lower":1126.9178840891268,"rating_upper":1138.617338485755,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.907894081892463,"vote_count":20781},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1126.9178840891268,"type":"confidence_interval","upper":1138.617338485755}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1132.767611287441,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d3736fe88ccaad8779bc","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4303f4a77031da3a1c52ae0a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1199.4437910533609,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a490fe89f0e2b4b57d4a82254d413e5b636cd71","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2500,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=13","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2500,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1199.4437910533609,"rating_lower":1194.2059273750187,"rating_upper":1204.681654731703,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.141874270550868,"vote_count":59909},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1194.2059273750187,"type":"confidence_interval","upper":1204.681654731703}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1199.4437910533609,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-43839f846cc2c5a68707","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5879cc04bd3ca6b06e7e03e2","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Arena search presentations of GPT-5.2; search and reasoning settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena search presentations of GPT-5.2; search and reasoning settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1206.9460713540655,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"64e8fe7347ccdbb05c467b9bc9675d2b38414f22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2497,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=10","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2497,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1206.9460713540655,"rating_lower":1201.349891438418,"rating_upper":1212.542251269713,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.152431435576114,"vote_count":52712},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1201.349891438418,"type":"confidence_interval","upper":1212.542251269713}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1206.9460713540655,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-612ee6482e9cdbb11317","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1c1b4ed72231ea0c7e8d2a07","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Arena search presentations of GPT-5.2; search and reasoning settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena search presentations of GPT-5.2; search and reasoning settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-search-non-reasoning","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1172.1381005441356,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"64e8fe7347ccdbb05c467b9bc9675d2b38414f22","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2506,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=19","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2506,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1172.1381005441356,"rating_lower":1166.5908120147346,"rating_upper":1177.6853890735365,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.010605206191979,"vote_count":75658},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1166.5908120147346,"type":"confidence_interval","upper":1177.6853890735365}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1172.1381005441356,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3a6ff4e45fc292a5297b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6a996e7720c320a1fd8c0f64","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena search endpoint presentation of GPT-5.4.","registry":"model_aliases.json"},"mappingNote":"Arena search endpoint presentation of GPT-5.4.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1197.1574686457088,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ff2b1c13b4a09fdcfd340258fc145f84dae27eb8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2502,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=15","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2502,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1197.1574686457088,"rating_lower":1191.7381114987943,"rating_upper":1202.5768257926236,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.645385062466708,"vote_count":110116},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1191.7381114987943,"type":"confidence_interval","upper":1202.5768257926236}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1197.1574686457088,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3121c90c8e08b20b5c29","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-464bf310c1b82f8210623022","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1242.052705813796,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"118977ff4bebe5dcecd7f366e9964436bfd78612","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2489,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=2","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2489,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1242.052705813796,"rating_lower":1236.7348808525694,"rating_upper":1247.3705307750226,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.361594549940995,"vote_count":89873},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1236.7348808525694,"type":"confidence_interval","upper":1247.3705307750226}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1242.052705813796,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-58fe311ccf9685baaf33","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8070d426b107e80ee37239be","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1257.2603500980206,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9cca71810d4cd58068cf3f1d952326934659dad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2487,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=0","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2487,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1257.2603500980206,"rating_lower":1249.7949622380377,"rating_upper":1264.7257379580033,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":14.508034187363197,"vote_count":29663},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.7949622380377,"type":"confidence_interval","upper":1264.7257379580033}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1257.2603500980206,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-359af77a88e30e4368d3","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4ec565bbd96b3bf02151d41a","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingEvidence":{"confidence":"high","note":"Arena search endpoint presentation of Grok 4 Fast.","registry":"model_aliases.json"},"mappingNote":"Arena search endpoint presentation of Grok 4 Fast.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1170.50892593241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23ecddb353b73b6677f545ae2631c8a9b64e19dd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2508,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=21","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2508,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1170.50892593241,"rating_lower":1165.9172483770526,"rating_upper":1175.100603487767,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.488410459795556,"vote_count":41794},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1165.9172483770526,"type":"confidence_interval","upper":1175.100603487767}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1170.50892593241,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-48107a62f45acae3750d","canonicalModelId":"xai/grok-4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8a08ce7f85de876f893e75cb","mappingCandidates":["xai/grok-4@current"],"mappingEvidence":{"confidence":"high","note":"Arena search endpoint presentation of Grok 4.","registry":"model_aliases.json"},"mappingNote":"Arena search endpoint presentation of Grok 4.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4","modelRef":"grok-4-search","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1141.5180851711918,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9738da24021dd2a133ce0f7252129e19e54981a6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2514,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=27","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2514,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1141.5180851711918,"rating_lower":1135.5589982646327,"rating_upper":1147.477172077751,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.244070655821425,"vote_count":19108},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1135.5589982646327,"type":"confidence_interval","upper":1147.477172077751}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1141.5180851711918,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"arena-search","benchmarkName":"Arena Search","benchmarkVersion":null,"benchmarkVersionHint":"arena-search@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f9faee29c7bcedcd7bf9","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-44330158b5ab52e827070332","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-24","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"search","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1212.646185108397,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e639dc50a9a958e3f9b866c7ee378fdfebaa43c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2494,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=search;split=latest;row_idx=7","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2494,"metadata":{"category":"overall","config":"search","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1212.646185108397,"rating_lower":1205.7834727411348,"rating_upper":1219.5088974756593,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.260139502748519,"vote_count":31505},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1205.7834727411348,"type":"confidence_interval","upper":1219.5088974756593}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1212.646185108397,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-723be4aa11a1186d739d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-565200f47136ca2e363ffedc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5.0,"rawValue":1494.684788253507,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99a1ecfa8a8bb90361f9cbe01c17903b9322b52e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":5,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=4","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":5,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1494.684788253507,"rating_lower":1489.5074693848928,"rating_upper":1499.862107122121,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.97772172459818,"vote_count":24331.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1489.5074693848928,"type":"confidence_interval","upper":1499.862107122121}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1494.684788253507,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-19ebbddf23134e683cc6","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4d73795515b66664dabd3106","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5.0,"rawValue":1519.2356478977404,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99a1ecfa8a8bb90361f9cbe01c17903b9322b52e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":763,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=762","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":763,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1519.2356478977404,"rating_lower":1511.1445567439062,"rating_upper":1527.3267390515748,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.041899735326904,"vote_count":6552.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1511.1445567439062,"type":"confidence_interval","upper":1527.3267390515748}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1519.2356478977404,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-88789f0826ecaa9e25b7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a49a6476a6ba1ea442b13cd4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":7.0,"rawValue":1544.239298121939,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99a1ecfa8a8bb90361f9cbe01c17903b9322b52e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":401,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=400","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":401,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1544.239298121939,"rating_lower":1527.917595003278,"rating_upper":1560.5610012406003,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":69.34813182393461,"vote_count":1498.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1527.917595003278,"type":"confidence_interval","upper":1560.5610012406003}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1544.239298121939,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b0a13f0f5e9a2099a437","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-26623cbd8e3cf1439abec5c7","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":82.0,"rawValue":1452.770283211219,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce1775094d22d60c680599299ce7a2c2e3904f8c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":840,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=839","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":840,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1452.770283211219,"rating_lower":1448.4350077927647,"rating_upper":1457.1055586296732,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.892571762739108,"vote_count":31146.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.4350077927647,"type":"confidence_interval","upper":1457.1055586296732}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1452.770283211219,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4e3b13f1ffae4b05f36b","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-46083043d86a2d2e9a68e838","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":143.0,"rawValue":1394.8817158702782,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce1775094d22d60c680599299ce7a2c2e3904f8c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":143,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=142","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":143,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1394.8817158702782,"rating_lower":1392.3170106337964,"rating_upper":1397.4464211067598,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":1.7122955775556243,"vote_count":122469.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1392.3170106337964,"type":"confidence_interval","upper":1397.4464211067598}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1394.8817158702782,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-efe8967d09e991d3d79b","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f19ff3eda0eb3f3e4d9c4098","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":144.0,"rawValue":1422.9434261424383,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce1775094d22d60c680599299ce7a2c2e3904f8c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":538,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=537","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":538,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1422.9434261424383,"rating_lower":1414.6277071381112,"rating_upper":1431.2591451467654,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.001281749111794,"vote_count":6421.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1414.6277071381112,"type":"confidence_interval","upper":1431.2591451467654}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1422.9434261424383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b2b18012865279ba2f4a","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ae688607720557f54688f08a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":153.0,"rawValue":1401.6403101454246,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"738761d3824edd077a506fd273ca741a4d26c285","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":911,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=910","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":911,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1401.6403101454246,"rating_lower":1394.4307515332207,"rating_upper":1408.8498687576284,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.530728249069508,"vote_count":7928.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.4307515332207,"type":"confidence_interval","upper":1408.8498687576284}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1401.6403101454246,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7aada8339ef7687fc535","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-07a03b229d31ff4b4c531591","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":167.0,"rawValue":1385.831066547008,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"738761d3824edd077a506fd273ca741a4d26c285","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":561,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=560","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":561,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1385.831066547008,"rating_lower":1373.5270108195596,"rating_upper":1398.1351222744568,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.409452088517725,"vote_count":2423.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1373.5270108195596,"type":"confidence_interval","upper":1398.1351222744568}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1385.831066547008,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ddfa8d08c137eef6fb9f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ed65d4db95824ba5f513b84f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":174.0,"rawValue":1364.290818279198,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"738761d3824edd077a506fd273ca741a4d26c285","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":174,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=173","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":174,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1364.290818279198,"rating_lower":1360.0060816531104,"rating_upper":1368.5755549052856,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.779165627400459,"vote_count":44326.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1360.0060816531104,"type":"confidence_interval","upper":1368.5755549052856}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1364.290818279198,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-21c1b00af31072596548","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d1a1e1c53f71d4a3ebaf3eb9","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":96.0,"rawValue":1442.6087338228988,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"738761d3824edd077a506fd273ca741a4d26c285","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":854,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=853","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":854,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1442.6087338228988,"rating_lower":1434.8123576252685,"rating_upper":1450.4051100205293,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.823020537805748,"vote_count":6707.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1434.8123576252685,"type":"confidence_interval","upper":1450.4051100205293}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1442.6087338228988,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-04f641f442b742c0d2b6","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-08dcc60a1364a553162d991c","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":158.0,"rawValue":1375.7212726765192,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"738761d3824edd077a506fd273ca741a4d26c285","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":158,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=157","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":158,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1375.7212726765192,"rating_lower":1371.3375543680843,"rating_upper":1380.1049909849542,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.002523026977365,"vote_count":36987.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1371.3375543680843,"type":"confidence_interval","upper":1380.1049909849542}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1375.7212726765192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0dd2aa8606f8e5aa8aaf","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c12e0775113004d87dfb27b2","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":168.0,"rawValue":1385.69489723648,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"738761d3824edd077a506fd273ca741a4d26c285","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":562,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=561","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":562,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1385.69489723648,"rating_lower":1371.7535399245025,"rating_upper":1399.6362545484576,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":50.59573791425167,"vote_count":1918.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1371.7535399245025,"type":"confidence_interval","upper":1399.6362545484576}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1385.69489723648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-39544c00647e3cc0ceb0","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-934c49bb50d96878a2767dc4","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":15.0,"rawValue":1498.4999714495004,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c00365d5050fede1f1ef2a75a1d5e7b01cf6e669","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":773,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=772","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":773,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1498.4999714495004,"rating_lower":1493.1048578959605,"rating_upper":1503.8950850030403,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.577134524724725,"vote_count":17320.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1493.1048578959605,"type":"confidence_interval","upper":1503.8950850030403}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1498.4999714495004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-defc2fd0dfa78904d6ff","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3fb626b844ded292d9debc57","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42.0,"rawValue":1449.8532784846016,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c00365d5050fede1f1ef2a75a1d5e7b01cf6e669","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":42,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=41","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":42,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1449.8532784846016,"rating_lower":1446.6886735662417,"rating_upper":1453.0178834029614,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":2.6070107104513647,"vote_count":70985.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1446.6886735662417,"type":"confidence_interval","upper":1453.0178834029614}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1449.8532784846016,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-efeb464e89ebd379595d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-159c38514aa5a59834b2d028","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":70.0,"rawValue":1473.9154377418088,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c00365d5050fede1f1ef2a75a1d5e7b01cf6e669","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":464,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=463","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":464,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1473.9154377418088,"rating_lower":1463.4938839713757,"rating_upper":1484.336991512242,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.272796367086283,"vote_count":3709.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1463.4938839713757,"type":"confidence_interval","upper":1484.336991512242}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1473.9154377418088,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9815aaf1cc91608d48cd","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-43131b87e65c86e9f1e5b1b9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12.0,"rawValue":1502.334287039247,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c00365d5050fede1f1ef2a75a1d5e7b01cf6e669","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":770,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=769","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":770,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1502.334287039247,"rating_lower":1495.0940458936766,"rating_upper":1509.5745281848174,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.646141815607773,"vote_count":7648.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1495.0940458936766,"type":"confidence_interval","upper":1509.5745281848174}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1502.334287039247,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-210d3f7e3689d8c7a6da","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e7f4985d9a7e979eca7405ea","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45.0,"rawValue":1446.6977054302,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c00365d5050fede1f1ef2a75a1d5e7b01cf6e669","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":45,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=44","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":45,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1446.6977054302,"rating_lower":1442.834790077404,"rating_upper":1450.560620782996,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.884491730714779,"vote_count":37198.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1442.834790077404,"type":"confidence_interval","upper":1450.560620782996}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1446.6977054302,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-61dc025c9427ee0cb451","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3dcbafb79e1decd664359400","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":93.0,"rawValue":1461.0518393877824,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c00365d5050fede1f1ef2a75a1d5e7b01cf6e669","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":487,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=486","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":487,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1461.0518393877824,"rating_lower":1447.263214791758,"rating_upper":1474.8404639838066,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":49.493220446844276,"vote_count":1878.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.263214791758,"type":"confidence_interval","upper":1474.8404639838066}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1461.0518393877824,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e06a2b32f404c3d5e48c","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-72ddbb43542525d82b165399","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":2.0,"rawValue":1533.8798441359052,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"025c8f30537475f60cf5eb0d5bfc4d684eba442e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":760,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=759","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":760,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1533.8798441359052,"rating_lower":1528.4649806440525,"rating_upper":1539.294707627758,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.632711426567258,"vote_count":21288.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1528.4649806440525,"type":"confidence_interval","upper":1539.294707627758}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1533.8798441359052,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b1cf638452380c65c53c","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3c598c07cd2694769b42625f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":4.0,"rawValue":1497.1950053437704,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"025c8f30537475f60cf5eb0d5bfc4d684eba442e","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":4,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=3","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":4,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1497.1950053437704,"rating_lower":1493.7907197528425,"rating_upper":1500.5992909346985,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.01686440634704,"vote_count":76362.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1493.7907197528425,"type":"confidence_interval","upper":1500.5992909346985}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1497.1950053437704,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b917ce2ecec75c06dea1","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-362bc5b596fdc66ba3cfbb7c","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5.0,"rawValue":1550.6343188146861,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"025c8f30537475f60cf5eb0d5bfc4d684eba442e","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":399,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=398","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":399,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1550.6343188146861,"rating_lower":1540.2465137036818,"rating_upper":1561.0221239256905,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.089978328782017,"vote_count":4201.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1540.2465137036818,"type":"confidence_interval","upper":1561.0221239256905}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1550.6343188146861,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bc851fe32db17c10d272","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-efd0eed446610099febd250d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":1.0,"rawValue":1535.4160363275219,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"025c8f30537475f60cf5eb0d5bfc4d684eba442e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":759,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=758","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":759,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1535.4160363275219,"rating_lower":1529.7676235899453,"rating_upper":1541.0644490650984,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.305325644035362,"vote_count":18844.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1529.7676235899453,"type":"confidence_interval","upper":1541.0644490650984}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1535.4160363275219,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b75444d9f6b80fa836d0","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5d68a4f3bbc38896ed4f9d4d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3.0,"rawValue":1502.6783117818993,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"025c8f30537475f60cf5eb0d5bfc4d684eba442e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":3,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=2","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":3,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1502.6783117818993,"rating_lower":1499.1840684930387,"rating_upper":1506.1725550707597,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.1784113097795825,"vote_count":72359.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1499.1840684930387,"type":"confidence_interval","upper":1506.1725550707597}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1502.6783117818993,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4e8536a0cc4bb5cee268","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fc3e03c69a1de251f0524ade","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":4.0,"rawValue":1554.5618817134211,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"025c8f30537475f60cf5eb0d5bfc4d684eba442e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":398,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=397","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":398,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1554.5618817134211,"rating_lower":1544.1263067332195,"rating_upper":1564.9974566936228,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.348924263031034,"vote_count":4196.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1544.1263067332195,"type":"confidence_interval","upper":1564.9974566936228}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1554.5618817134211,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b92e20ed1074663bd0ac","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2a3f29f4c525c0fb53e39877","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":7.0,"rawValue":1515.6750721273409,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f19614851d79b91cd74541ddbf42a47d15932d7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":765,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=764","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":765,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1515.6750721273409,"rating_lower":1509.6799776752791,"rating_upper":1521.6701665794026,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.356122027268285,"vote_count":17362.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1509.6799776752791,"type":"confidence_interval","upper":1521.6701665794026}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1515.6750721273409,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7dac9ed8caf274893eb5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-42c37695164fbd21f8b13c87","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9.0,"rawValue":1482.7423547986384,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f19614851d79b91cd74541ddbf42a47d15932d7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":9,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=8","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":9,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1482.7423547986384,"rating_lower":1478.8709960612491,"rating_upper":1486.6137135360277,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.901491379478716,"vote_count":61304.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1478.8709960612491,"type":"confidence_interval","upper":1486.6137135360277}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1482.7423547986384,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b4e03c5a21db923004a1","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-25b5e287a2ed3c3e5707ebae","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":19.0,"rawValue":1528.1108979767837,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f19614851d79b91cd74541ddbf42a47d15932d7","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":413,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=412","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":413,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1528.1108979767837,"rating_lower":1516.9025753811447,"rating_upper":1539.3192205724226,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.70280934189751,"vote_count":3434.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1516.9025753811447,"type":"confidence_interval","upper":1539.3192205724226}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1528.1108979767837,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-69596b5863107681fe1f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1eeb61d4e8c05e901da04dc8","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":6.0,"rawValue":1519.1148988655398,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f19614851d79b91cd74541ddbf42a47d15932d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":764,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=763","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":764,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1519.1148988655398,"rating_lower":1513.0860172833623,"rating_upper":1525.143780447717,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.46187759090691,"vote_count":17212.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1513.0860172833623,"type":"confidence_interval","upper":1525.143780447717}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1519.1148988655398,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f2ad16383852584ec70c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ca0d0d4b1bed486d3d53895b","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":7.0,"rawValue":1490.1204074818936,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f19614851d79b91cd74541ddbf42a47d15932d7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":7,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=6","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":7,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1490.1204074818936,"rating_lower":1486.249031257029,"rating_upper":1493.9917837067585,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.9015266267359454,"vote_count":60203.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1486.249031257029,"type":"confidence_interval","upper":1493.9917837067585}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1490.1204074818936,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-eb1074d8f9bf1ba1f898","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-00f82ba42564ac839c99b284","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12.0,"rawValue":1534.9680371881902,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f19614851d79b91cd74541ddbf42a47d15932d7","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":406,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=405","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":406,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1534.9680371881902,"rating_lower":1523.7628761926924,"rating_upper":1546.1731981836883,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.684362580865375,"vote_count":3382.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1523.7628761926924,"type":"confidence_interval","upper":1546.1731981836883}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1534.9680371881902,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-51e68b7bcb3519ee9366","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b05e539994a1c0ba73efccdf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":34.0,"rawValue":1484.3170781766582,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"424ae8e20ed61e07b9dd93f06dbd7aeb18b59b06","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":792,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=791","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":792,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1484.3170781766582,"rating_lower":1477.7326774169119,"rating_upper":1490.9014789364046,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.285903451937674,"vote_count":12737.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1477.7326774169119,"type":"confidence_interval","upper":1490.9014789364046}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1484.3170781766582,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-13aeabddc3f01a41d826","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-915cda4455cdadd9cf5eb6e0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":36.0,"rawValue":1505.2729249443958,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"424ae8e20ed61e07b9dd93f06dbd7aeb18b59b06","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":430,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=429","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":430,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1505.2729249443958,"rating_lower":1492.6384363383359,"rating_upper":1517.9074135504554,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.55460458842358,"vote_count":2565.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1492.6384363383359,"type":"confidence_interval","upper":1517.9074135504554}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1505.2729249443958,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4d95994d370cb238ca0a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8077b814799e387eddaf3ce6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":38.0,"rawValue":1451.946373879298,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"424ae8e20ed61e07b9dd93f06dbd7aeb18b59b06","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":38,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=37","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":38,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1451.946373879298,"rating_lower":1447.5134778549452,"rating_upper":1456.379269903651,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.115391855007104,"vote_count":45331.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.5134778549452,"type":"confidence_interval","upper":1456.379269903651}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1451.946373879298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-974f0a67713032f31b1d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-82cc33023b255da2fca0525d","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":24.0,"rawValue":1489.6262287811883,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"424ae8e20ed61e07b9dd93f06dbd7aeb18b59b06","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":782,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=781","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":782,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1489.6262287811883,"rating_lower":1483.0060265110517,"rating_upper":1496.2464310513249,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.40896730726955,"vote_count":12425.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1483.0060265110517,"type":"confidence_interval","upper":1496.2464310513249}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1489.6262287811883,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-29a3b027222290747fc5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-af66a9a956dd795efd57e2ae","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":30.0,"rawValue":1460.9095664685228,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"424ae8e20ed61e07b9dd93f06dbd7aeb18b59b06","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":30,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=29","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":30,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1460.9095664685228,"rating_lower":1456.5054558203676,"rating_upper":1465.3136771166778,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.049173115355268,"vote_count":44676.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1456.5054558203676,"type":"confidence_interval","upper":1465.3136771166778}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1460.9095664685228,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-571faa23aa832fee5bcb","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-86b47fb3c0c7a3b5f13f36ca","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":41.0,"rawValue":1501.446206226374,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"424ae8e20ed61e07b9dd93f06dbd7aeb18b59b06","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":435,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=434","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":435,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1501.446206226374,"rating_lower":1488.8073909608368,"rating_upper":1514.085021491911,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.583070071153024,"vote_count":2555.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1488.8073909608368,"type":"confidence_interval","upper":1514.085021491911}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1501.446206226374,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4e3be9070d23890fe899","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ce323b75362323b431e08cf4","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1.0,"rawValue":1504.2025661543232,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f93a56f182bed6a102d3c985b86098f7ec975ae0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=0","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1504.2025661543232,"rating_lower":1499.4103052358735,"rating_upper":1508.994827072773,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.9783966931478565,"vote_count":27610.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1499.4103052358735,"type":"confidence_interval","upper":1508.994827072773}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1504.2025661543232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e4ddee18f18a259a2ef1","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7194ca77d2c320b080ee4b92","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2.0,"rawValue":1585.0807825560232,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f93a56f182bed6a102d3c985b86098f7ec975ae0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":396,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=395","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":396,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1585.0807825560232,"rating_lower":1570.273004441975,"rating_upper":1599.8885606700715,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.07995397312716,"vote_count":1939.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1570.273004441975,"type":"confidence_interval","upper":1599.8885606700715}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1585.0807825560232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-105a86074b0fbd5d0c7c","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fc71e936af09b840ae5d92e5","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":3.0,"rawValue":1531.5185073420098,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f93a56f182bed6a102d3c985b86098f7ec975ae0","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":761,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=760","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":761,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1531.5185073420098,"rating_lower":1523.5498817308835,"rating_upper":1539.4871329531359,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.529916652555023,"vote_count":7416.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1523.5498817308835,"type":"confidence_interval","upper":1539.4871329531359}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1531.5185073420098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4eed6d2d77a75dc504cd","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f058636aceda8a0937217553","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1.0,"rawValue":1588.2900384300049,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f93a56f182bed6a102d3c985b86098f7ec975ae0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":395,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=394","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":395,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1588.2900384300049,"rating_lower":1567.6427018659879,"rating_upper":1608.937374994022,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":110.97672188784877,"vote_count":949.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1567.6427018659879,"type":"confidence_interval","upper":1608.937374994022}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1588.2900384300049,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4bd71591a02a0cec38e7","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-025b1677857c3f8d24e5f7b9","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2.0,"rawValue":1504.0374724478093,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f93a56f182bed6a102d3c985b86098f7ec975ae0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=1","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1504.0374724478093,"rating_lower":1497.7236012192109,"rating_upper":1510.3513436764074,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.377560128086495,"vote_count":13403.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1497.7236012192109,"type":"confidence_interval","upper":1510.3513436764074}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1504.0374724478093,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b65ec7a75b943c286975","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6f180f0c11efaff98f0ddaf6","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4.0,"rawValue":1528.395960285115,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f93a56f182bed6a102d3c985b86098f7ec975ae0","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":762,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=761","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":762,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1528.395960285115,"rating_lower":1517.9605735944426,"rating_upper":1538.8313469757875,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.347901270535303,"vote_count":3570.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1517.9605735944426,"type":"confidence_interval","upper":1538.8313469757875}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1528.395960285115,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e612f086ace01c2adebe","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-641888df0bdb73508d75ce61","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32.0,"rawValue":1485.4495918725258,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31841914744a0959a24b09899d9abd6691ccc34","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":790,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=789","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":790,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1485.4495918725258,"rating_lower":1480.3647439140145,"rating_upper":1490.5344398310374,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.730692679013569,"vote_count":19145.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1480.3647439140145,"type":"confidence_interval","upper":1490.5344398310374}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1485.4495918725258,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-aaa0a69415787eb00de4","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-93142edb4299a5096ca4d289","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":68.0,"rawValue":1437.9788159129344,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31841914744a0959a24b09899d9abd6691ccc34","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":68,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=67","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":68,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1437.9788159129344,"rating_lower":1435.0854981582984,"rating_upper":1440.8721336675706,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":2.179194941306141,"vote_count":80858.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1435.0854981582984,"type":"confidence_interval","upper":1440.8721336675706}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1437.9788159129344,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8d3e1d1e955bfb440f71","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c5a05e48e56e4984bf94ed02","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":86.0,"rawValue":1466.401517747521,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31841914744a0959a24b09899d9abd6691ccc34","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":480,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=479","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":480,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1466.401517747521,"rating_lower":1456.302309345107,"rating_upper":1476.500726149935,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":26.550853495017044,"vote_count":3945.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1456.302309345107,"type":"confidence_interval","upper":1476.500726149935}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1466.401517747521,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d153ac82f6e03cfc8402","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f8b61cac5d32b76b1bce5246","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26.0,"rawValue":1488.2099563202128,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31841914744a0959a24b09899d9abd6691ccc34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":784,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=783","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":784,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1488.2099563202128,"rating_lower":1483.2800160358777,"rating_upper":1493.1398966045476,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.326844134363792,"vote_count":19352.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1483.2800160358777,"type":"confidence_interval","upper":1493.1398966045476}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1488.2099563202128,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-66ad57fbd1f3d905b0a8","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c8fae60c2be1e049937019a0","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":78.0,"rawValue":1433.3331108484715,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31841914744a0959a24b09899d9abd6691ccc34","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":78,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=77","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":78,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1433.3331108484715,"rating_lower":1430.5411363124103,"rating_upper":1436.1250853845327,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":2.029208739143822,"vote_count":82473.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.5411363124103,"type":"confidence_interval","upper":1436.1250853845327}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1433.3331108484715,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f638608f588ba8f29cae","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ddfc58e95078025ad9dc5f32","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":83.0,"rawValue":1467.1058479923634,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31841914744a0959a24b09899d9abd6691ccc34","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":477,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=476","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":477,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1467.1058479923634,"rating_lower":1457.1041137970358,"rating_upper":1477.1075821876912,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":26.04080678285377,"vote_count":3932.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.1041137970358,"type":"confidence_interval","upper":1477.1075821876912}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1467.1058479923634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7986c0181510d89dc86a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4aa491b283cc2441a56eeed6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":10.0,"rawValue":1503.7881653396375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a3f61374af6352b09e05b3e4230d42c55ed9ab7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":768,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=767","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":768,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1503.7881653396375,"rating_lower":1498.1091067296172,"rating_upper":1509.467223949658,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.395692418282703,"vote_count":18416.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1498.1091067296172,"type":"confidence_interval","upper":1509.467223949658}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1503.7881653396375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d890c49ef6d76161c1c8","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9408ab2bc22ddde1d591c100","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31.0,"rawValue":1457.5719381664967,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a3f61374af6352b09e05b3e4230d42c55ed9ab7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":31,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=30","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":31,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1457.5719381664967,"rating_lower":1453.9762686277338,"rating_upper":1461.1676077052593,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.365606670658077,"vote_count":66512.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1453.9762686277338,"type":"confidence_interval","upper":1461.1676077052593}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1457.5719381664967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf615eda0ed2a4d60b31","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f420cdd350a8fa9f9f147244","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":51.0,"rawValue":1492.7292671098887,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a3f61374af6352b09e05b3e4230d42c55ed9ab7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":445,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=444","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":445,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1492.7292671098887,"rating_lower":1481.793746999446,"rating_upper":1503.6647872203312,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.13025693303796,"vote_count":3548.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1481.793746999446,"type":"confidence_interval","upper":1503.6647872203312}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1492.7292671098887,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5b4dd2c3ffa281198e95","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-01ed006de364bf1b453f1120","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":31.0,"rawValue":1485.463794561637,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de244f86ef116adb7e74366d73b39c8b8c26d5b2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":789,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=788","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":789,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1485.463794561637,"rating_lower":1477.8433190522924,"rating_upper":1493.084270070982,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.117081738762735,"vote_count":7626.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1477.8433190522924,"type":"confidence_interval","upper":1493.084270070982}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1485.463794561637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-89edd50871da75ef7b4c","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-76f6286e17d2c4a25002509b","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":57.0,"rawValue":1442.2497390160604,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de244f86ef116adb7e74366d73b39c8b8c26d5b2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":57,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=56","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":57,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1442.2497390160604,"rating_lower":1437.3391835024922,"rating_upper":1447.1602945296286,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.2771870212258865,"vote_count":27366.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1437.3391835024922,"type":"confidence_interval","upper":1447.1602945296286}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1442.2497390160604,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-91c94a0257188598cff4","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2b2b03f55fcbaf0a6bb08407","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":60.0,"rawValue":1482.3095417342138,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de244f86ef116adb7e74366d73b39c8b8c26d5b2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":454,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=453","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":454,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1482.3095417342138,"rating_lower":1467.5019800816622,"rating_upper":1497.1171033867652,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.07828518502687,"vote_count":1805.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1467.5019800816622,"type":"confidence_interval","upper":1497.1171033867652}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1482.3095417342138,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4dad53bcc427d6b6f166","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9f5d591e0c16975be8032b60","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek-R1-0528","modelRef":"deepseek-r1-0528","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":85.0,"rawValue":1427.9090691139086,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e101851b8071266a149aeeb3f38ff19fae5d80","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":85,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=84","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":85,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1427.9090691139086,"rating_lower":1422.2839388208447,"rating_upper":1433.5341994069722,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.236998570305262,"vote_count":18492.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1422.2839388208447,"type":"confidence_interval","upper":1433.5341994069722}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1427.9090691139086,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2f096c7dd362b8bb1e77","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-877652b38949a49d7885f0ad","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek-R1-0528","modelRef":"deepseek-r1-0528","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":102.0,"rawValue":1456.7251609405107,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e101851b8071266a149aeeb3f38ff19fae5d80","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":496,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=495","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":496,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1456.7251609405107,"rating_lower":1437.4113141642185,"rating_upper":1476.0390077168029,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":97.1049527561218,"vote_count":1013.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1437.4113141642185,"type":"confidence_interval","upper":1476.0390077168029}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1456.7251609405107,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8f0ed56c4738da52e298","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7a32b09bf13a6e713c88e72","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek-R1-0528","modelRef":"deepseek-r1-0528","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":122.0,"rawValue":1427.0808103160361,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e101851b8071266a149aeeb3f38ff19fae5d80","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":880,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=879","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":880,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1427.0808103160361,"rating_lower":1415.7704379509537,"rating_upper":1438.3911826811188,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":33.30102677339359,"vote_count":2733.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1415.7704379509537,"type":"confidence_interval","upper":1438.3911826811188}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1427.0808103160361,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8f98a8ae76bcc1707067","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-55434b64fe5970ef4f6b5f2c","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1","modelRef":"deepseek-v3.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":104.0,"rawValue":1455.8103764607301,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b94deb9e6093047a4e59e117bf10a58cd2797a41","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":498,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=497","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":498,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1455.8103764607301,"rating_lower":1437.9835707945997,"rating_upper":1473.6371821268606,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":82.72768630136035,"vote_count":1118.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1437.9835707945997,"type":"confidence_interval","upper":1473.6371821268606}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1455.8103764607301,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b3d563ffcb6be5f9e122","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-55ac59e3cc366f57f11b5eab","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1","modelRef":"deepseek-v3.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":105.0,"rawValue":1419.1240636772848,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b94deb9e6093047a4e59e117bf10a58cd2797a41","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":105,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=104","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":105,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1419.1240636772848,"rating_lower":1413.1432886149626,"rating_upper":1425.104838739607,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.311480876328043,"vote_count":15002.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1413.1432886149626,"type":"confidence_interval","upper":1425.104838739607}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1419.1240636772848,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-83b49f88d56bc2e5daf0","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-68a67546179f1603804b79d7","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1","modelRef":"deepseek-v3.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":135.0,"rawValue":1414.748302689449,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b94deb9e6093047a4e59e117bf10a58cd2797a41","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":893,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=892","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":893,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1414.748302689449,"rating_lower":1403.3092008230283,"rating_upper":1426.18740455587,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.06337477976881,"vote_count":2640.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1403.3092008230283,"type":"confidence_interval","upper":1426.18740455587}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1414.748302689449,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47b7090c8953877277a0","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b9c2b7775022246baa75f161","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1 Terminus","modelRef":"deepseek-v3.1-terminus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":112.0,"rawValue":1416.6386560476492,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"331b498e018c7ff8826f70c0459a715b0f7ec577","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":112,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=111","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":112,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1416.6386560476492,"rating_lower":1407.0377387972612,"rating_upper":1426.2395732980372,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.995470562441525,"vote_count":3700.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1407.0377387972612,"type":"confidence_interval","upper":1426.2395732980372}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1416.6386560476492,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-37bad79acd89e3cba977","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9cef733c8963501d396fb16f","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1 Terminus","modelRef":"deepseek-v3.1-terminus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":148.0,"rawValue":1406.5948738610255,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"331b498e018c7ff8826f70c0459a715b0f7ec577","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":906,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=905","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":906,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1406.5948738610255,"rating_lower":1386.1031170301403,"rating_upper":1427.0866306919104,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":109.3105816347779,"vote_count":779.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1386.1031170301403,"type":"confidence_interval","upper":1427.0866306919104}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1406.5948738610255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-52f4c24e34faa6a2eb11","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6985f355e39c9663c64c3d3d","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1 Terminus","modelRef":"deepseek-v3.1-terminus-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":102.0,"rawValue":1419.5842644243733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"331b498e018c7ff8826f70c0459a715b0f7ec577","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":102,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=101","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":102,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1419.5842644243733,"rating_lower":1409.6282762654123,"rating_upper":1429.5402525833342,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.803140121507,"vote_count":3470.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1409.6282762654123,"type":"confidence_interval","upper":1429.5402525833342}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1419.5842644243733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1e2ac8b563bbc34d51e0","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8e1bc949df10eca0f45d2453","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1 Terminus","modelRef":"deepseek-v3.1-terminus-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":119.0,"rawValue":1428.5142413891792,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"331b498e018c7ff8826f70c0459a715b0f7ec577","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":877,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=876","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":877,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1428.5142413891792,"rating_lower":1405.2812900529402,"rating_upper":1451.7471927254182,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":140.51173082587326,"vote_count":641.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1405.2812900529402,"type":"confidence_interval","upper":1451.7471927254182}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1428.5142413891792,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-925ef9651663b1636fca","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3ad657921e7ebbf4d3afee21","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1","modelRef":"deepseek-v3.1-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":75.0,"rawValue":1470.4827142810077,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b94deb9e6093047a4e59e117bf10a58cd2797a41","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":469,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=468","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":469,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1470.4827142810077,"rating_lower":1449.9637206896725,"rating_upper":1491.0017078723429,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":109.60135658181507,"vote_count":883.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1449.9637206896725,"type":"confidence_interval","upper":1491.0017078723429}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1470.4827142810077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2cd3970ab7b5af99afbc","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c68e9a12999e78898cd7b7f6","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1","modelRef":"deepseek-v3.1-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":113.0,"rawValue":1416.5083613666134,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b94deb9e6093047a4e59e117bf10a58cd2797a41","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":113,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=112","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":113,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1416.5083613666134,"rating_lower":1409.939700454049,"rating_upper":1423.0770222791778,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.232010597592433,"vote_count":11786.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1409.939700454049,"type":"confidence_interval","upper":1423.0770222791778}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1416.5083613666134,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47d764acd51a437cd443","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-68145544e60fa37a8f76eeec","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.1","modelRef":"deepseek-v3.1-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":131.0,"rawValue":1418.4119018236506,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b94deb9e6093047a4e59e117bf10a58cd2797a41","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":889,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=888","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":889,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1418.4119018236506,"rating_lower":1405.0763431981259,"rating_upper":1431.7474604491754,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":46.29416379443926,"vote_count":1911.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1405.0763431981259,"type":"confidence_interval","upper":1431.7474604491754}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1418.4119018236506,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4561fe39928f8accf77f","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6fd7aeac36e0e211ac94608a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":86.0,"rawValue":1449.0826070625546,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ca39a3a6fcef840063bc3810fa377177627664b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":844,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=843","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":844,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1449.0826070625546,"rating_lower":1442.7338011352049,"rating_upper":1455.4314129899044,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.4927160707837,"vote_count":10642.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1442.7338011352049,"type":"confidence_interval","upper":1455.4314129899044}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1449.0826070625546,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4dcd64d1ce5e7746875d","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-411f073f4bad4eaef70314d5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":90.0,"rawValue":1424.6209587565081,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ca39a3a6fcef840063bc3810fa377177627664b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":90,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=89","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":90,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1424.6209587565081,"rating_lower":1421.1085492765558,"rating_upper":1428.1333682364602,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.2115456472937445,"vote_count":47295.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1421.1085492765558,"type":"confidence_interval","upper":1428.1333682364602}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1424.6209587565081,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c0ad538ab246e2dfdbfe","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f37589e083103561ae884011","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":96.0,"rawValue":1459.500925724597,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ca39a3a6fcef840063bc3810fa377177627664b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":490,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=489","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":490,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1459.500925724597,"rating_lower":1447.025691621072,"rating_upper":1471.976159828122,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.51363640796062,"vote_count":2497.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.025691621072,"type":"confidence_interval","upper":1471.976159828122}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1459.500925724597,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-693617aa379a1d597009","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-690c6e7af75562e94f9ddb70","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":89.0,"rawValue":1464.8964135646193,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a25ca90705fe61b6b9fb12280b877b0a91f1e3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":483,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=482","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":483,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1464.8964135646193,"rating_lower":1441.95758938053,"rating_upper":1487.8352377487086,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":136.97651842939223,"vote_count":642.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1441.95758938053,"type":"confidence_interval","upper":1487.8352377487086}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.8964135646193,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4e531d04a8f09c0ab50f","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8f11bae78f0c00350144c336","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":93.0,"rawValue":1422.5513334398902,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a25ca90705fe61b6b9fb12280b877b0a91f1e3e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":93,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=92","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":93,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1422.5513334398902,"rating_lower":1416.2052281557528,"rating_upper":1428.8974387240273,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.483791225458663,"vote_count":11982.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1416.2052281557528,"type":"confidence_interval","upper":1428.8974387240273}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1422.5513334398902,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ee186d959de6a34903e8","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-62e5ea5a0decc3f8fbae1e61","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":113.0,"rawValue":1433.403071405,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a25ca90705fe61b6b9fb12280b877b0a91f1e3e","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":871,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=870","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":871,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1433.403071405,"rating_lower":1421.7631196926695,"rating_upper":1445.0430231173304,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.27005811841619,"vote_count":2501.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1421.7631196926695,"type":"confidence_interval","upper":1445.0430231173304}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1433.403071405,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d1c6e9c9513e3fea5a43","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bf58c6356f5f047a20a04cf9","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":91.0,"rawValue":1424.3777108565635,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a25ca90705fe61b6b9fb12280b877b0a91f1e3e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":91,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=90","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":91,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1424.3777108565635,"rating_lower":1417.8585560680717,"rating_upper":1430.8968656450552,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.063343677503111,"vote_count":9109.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1417.8585560680717,"type":"confidence_interval","upper":1430.8968656450552}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1424.3777108565635,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b6fa8155b9f672f08ec7","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fd771c59e87c584d83fa4076","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":95.0,"rawValue":1459.9497966396993,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a25ca90705fe61b6b9fb12280b877b0a91f1e3e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":489,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=488","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":489,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1459.9497966396993,"rating_lower":1427.09397360274,"rating_upper":1492.8056196766586,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":281.0143640277048,"vote_count":313.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1427.09397360274,"type":"confidence_interval","upper":1492.8056196766586}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1459.9497966396993,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf5948f3272648c664f2","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-47585a8666b61e994259f99f","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":104.0,"rawValue":1438.6471525701834,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a25ca90705fe61b6b9fb12280b877b0a91f1e3e","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":862,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=861","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":862,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1438.6471525701834,"rating_lower":1425.3750174988081,"rating_upper":1451.9192876415586,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":45.854863366985,"vote_count":1925.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1425.3750174988081,"type":"confidence_interval","upper":1451.9192876415586}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1438.6471525701834,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-35a2578e855dacde3507","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fb5332cac6fe50d0ccd2674b","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":80.0,"rawValue":1453.271826663273,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ca39a3a6fcef840063bc3810fa377177627664b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":838,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=837","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":838,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1453.271826663273,"rating_lower":1446.4954890906092,"rating_upper":1460.0481642359366,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.95346690984361,"vote_count":8556.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1446.4954890906092,"type":"confidence_interval","upper":1460.0481642359366}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1453.271826663273,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e3cc7c7acf81e1d248bc","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c003bd58ccffad0baef4b233","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":101.0,"rawValue":1419.8045027712699,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ca39a3a6fcef840063bc3810fa377177627664b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":101,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=100","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":101,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1419.8045027712699,"rating_lower":1416.1953849763543,"rating_upper":1423.4136205661855,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.390829334784708,"vote_count":41123.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1416.1953849763543,"type":"confidence_interval","upper":1423.4136205661855}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1419.8045027712699,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9ea904cd005f83940c21","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ec079d7853253e47b14f0a8e","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":103.0,"rawValue":1456.661947443516,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ca39a3a6fcef840063bc3810fa377177627664b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":497,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=496","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":497,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1456.661947443516,"rating_lower":1443.2391862384222,"rating_upper":1470.0847086486099,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":46.90158785469328,"vote_count":2009.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1443.2391862384222,"type":"confidence_interval","upper":1470.0847086486099}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1456.661947443516,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0d939138829346278a2b","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a2d2885daa92fd7244336d1a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":75.0,"rawValue":1455.186846872624,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5e46dd3570884ede1a7ca5fa45bacd28cf7fbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":833,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=832","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":833,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1455.186846872624,"rating_lower":1448.8477872774902,"rating_upper":1461.5259064677582,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.460525135448458,"vote_count":14374.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.8477872774902,"type":"confidence_interval","upper":1461.5259064677582}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1455.186846872624,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1d88d948d6482dcf1e66","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3e0a5d5ef4cebf7c700f93e6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":79.0,"rawValue":1431.6067430969854,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5e46dd3570884ede1a7ca5fa45bacd28cf7fbb","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":79,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=78","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":79,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1431.6067430969854,"rating_lower":1427.46894300163,"rating_upper":1435.7445431923409,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.457001995411222,"vote_count":49001.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1427.46894300163,"type":"confidence_interval","upper":1435.7445431923409}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1431.6067430969854,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3e8954d69296ea5ae8a6","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-40f90948a1b72b9df334324e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":85.0,"rawValue":1466.6123503314598,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5e46dd3570884ede1a7ca5fa45bacd28cf7fbb","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":479,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=478","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":479,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1466.6123503314598,"rating_lower":1453.952636180831,"rating_upper":1479.2720644820886,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.72070295593381,"vote_count":2529.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1453.952636180831,"type":"confidence_interval","upper":1479.2720644820886}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1466.6123503314598,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7089f23331ab156745b5","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-45d4161fafd19e6e4bdd4adb","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":79.0,"rawValue":1468.62205505987,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5e46dd3570884ede1a7ca5fa45bacd28cf7fbb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":473,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=472","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":473,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1468.62205505987,"rating_lower":1455.965812312426,"rating_upper":1481.2782978073144,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.69782573727829,"vote_count":2504.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1455.965812312426,"type":"confidence_interval","upper":1481.2782978073144}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1468.62205505987,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6af3e0df862b034b7f2b","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7323e6ab389102847a54fcc","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":92.0,"rawValue":1423.9732661575526,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5e46dd3570884ede1a7ca5fa45bacd28cf7fbb","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":92,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=91","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":92,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1423.9732661575526,"rating_lower":1419.7735325005583,"rating_upper":1428.172999814547,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.591423105898136,"vote_count":48679.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1419.7735325005583,"type":"confidence_interval","upper":1428.172999814547}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1423.9732661575526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ae857dfe9a3ff06c79da","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dd4b8e012760d4cc28e0c3b4","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":101.0,"rawValue":1439.749595500906,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5e46dd3570884ede1a7ca5fa45bacd28cf7fbb","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":859,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=858","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":859,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1439.749595500906,"rating_lower":1433.3629342694794,"rating_upper":1446.1362567323326,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.61821656535089,"vote_count":14320.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1433.3629342694794,"type":"confidence_interval","upper":1446.1362567323326}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1439.749595500906,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e43063ba8af4c75c492f","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f988ba94f5056000664c5e7f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40.0,"rawValue":1450.8668626687377,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b85b47cb70daa668846bd715e754d00bfce45ed9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":40,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=39","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":40,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1450.8668626687377,"rating_lower":1446.8703117522925,"rating_upper":1454.8634135851826,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.157904580857087,"vote_count":54263.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1446.8703117522925,"type":"confidence_interval","upper":1454.8634135851826}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1450.8668626687377,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-326de86d0882bee447d0","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b8b5227669dd4a5f2f8b61cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":55.0,"rawValue":1470.4607075830108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b85b47cb70daa668846bd715e754d00bfce45ed9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":813,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=812","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":813,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1470.4607075830108,"rating_lower":1464.4092698278491,"rating_upper":1476.5121453381723,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.53281048005044,"vote_count":16202.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1464.4092698278491,"type":"confidence_interval","upper":1476.5121453381723}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1470.4607075830108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-410bdcfecb0482078b8a","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5c8702185cacc3656f9645f2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":55.0,"rawValue":1486.8871828117715,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b85b47cb70daa668846bd715e754d00bfce45ed9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":449,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=448","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":449,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1486.8871828117715,"rating_lower":1474.821477919051,"rating_upper":1498.952887704492,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":37.897382570903744,"vote_count":2857.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1474.821477919051,"type":"confidence_interval","upper":1498.952887704492}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1486.8871828117715,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-25b7cc9b26a1647daa74","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1c2e719e8b7c2e625406c7b3","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53.0,"rawValue":1472.1368816979355,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5dba374548d1beab3ea24b95239e382ab750a99f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":811,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=810","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":811,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1472.1368816979355,"rating_lower":1454.2543010465254,"rating_upper":1490.0194623493458,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":83.2461587331054,"vote_count":1064.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1454.2543010465254,"type":"confidence_interval","upper":1490.0194623493458}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1472.1368816979355,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-988dc9ffd415096cfb09","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-61f2e3f4655654084d677ba4","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":58.0,"rawValue":1484.4183615633667,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5dba374548d1beab3ea24b95239e382ab750a99f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":452,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=451","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":452,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1484.4183615633667,"rating_lower":1445.730204824198,"rating_upper":1523.1065183025355,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":389.6367348287838,"vote_count":232.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1445.730204824198,"type":"confidence_interval","upper":1523.1065183025355}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1484.4183615633667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-76c1b4c2587b39e82759","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b08151d26da31dd6ced12cdf","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":64.0,"rawValue":1439.1643609270914,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5dba374548d1beab3ea24b95239e382ab750a99f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":64,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=63","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":64,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1439.1643609270914,"rating_lower":1429.315579626513,"rating_upper":1449.0131422276697,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.25043157669295,"vote_count":3621.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1429.315579626513,"type":"confidence_interval","upper":1449.0131422276697}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1439.1643609270914,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-464bac8d16b47a4865be","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2e21cfa386f45e3d0cd4ee93","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":50.0,"rawValue":1444.9973961807061,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b85b47cb70daa668846bd715e754d00bfce45ed9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":50,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=49","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":50,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1444.9973961807061,"rating_lower":1440.9310499302255,"rating_upper":1449.063742431187,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.304399083942609,"vote_count":51642.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1440.9310499302255,"type":"confidence_interval","upper":1449.063742431187}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1444.9973961807061,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-22955e475635f5d739ca","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3470a0222e0fcfd070765cd7","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":57.0,"rawValue":1484.7833319190365,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b85b47cb70daa668846bd715e754d00bfce45ed9","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":451,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=450","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":451,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1484.7833319190365,"rating_lower":1472.4470856551866,"rating_upper":1497.1195781828867,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.61593211998975,"vote_count":2760.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1472.4470856551866,"type":"confidence_interval","upper":1497.1195781828867}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1484.7833319190365,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cb5c6424b32755ce542b","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-efefe0b8e0841cd2c95e35cb","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":77.0,"rawValue":1453.6216111078786,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b85b47cb70daa668846bd715e754d00bfce45ed9","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":835,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=834","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":835,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1453.6216111078786,"rating_lower":1447.4279659798126,"rating_upper":1459.8152562359446,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.98611250646812,"vote_count":15128.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.4279659798126,"type":"confidence_interval","upper":1459.8152562359446}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1453.6216111078786,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3c72f90f2c94b8211ebc","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c38c7e0d690399494980a06a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":111.0,"rawValue":1417.2991182812218,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"abb0a77c3d5b700c69029b7cd58c195df30f8369","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":111,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=110","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":111,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1417.2991182812218,"rating_lower":1414.8768935709372,"rating_upper":1419.7213429915064,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":1.5273292832156191,"vote_count":124731.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1414.8768935709372,"type":"confidence_interval","upper":1419.7213429915064}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1417.2991182812218,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-24cda7b40a5d4b810307","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-80279be55a468b3d555a14be","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":112.0,"rawValue":1450.9718021708911,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"abb0a77c3d5b700c69029b7cd58c195df30f8369","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":506,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=505","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":506,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1450.9718021708911,"rating_lower":1443.1302630400999,"rating_upper":1458.8133413016824,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.006871037763474,"vote_count":6493.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1443.1302630400999,"type":"confidence_interval","upper":1458.8133413016824}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1450.9718021708911,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-12eb219f92c44f7f397e","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-abf046af69e4600ffb328663","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":127.0,"rawValue":1423.5268495127216,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"abb0a77c3d5b700c69029b7cd58c195df30f8369","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":885,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=884","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":885,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1423.5268495127216,"rating_lower":1419.2840022648195,"rating_upper":1427.769696760624,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.686176166214713,"vote_count":25967.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1419.2840022648195,"type":"confidence_interval","upper":1427.769696760624}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1423.5268495127216,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-303c0646f9a0c594fd19","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7b6382051ac174ed36c0a5e6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":32.0,"rawValue":1457.3391897717306,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb60abb91d9cdc5d1cc20128e2160a0b8b7b780d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":32,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=31","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":32,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1457.3391897717306,"rating_lower":1454.8823329091595,"rating_upper":1459.7960466343015,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":1.5713159830441765,"vote_count":124807.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1454.8823329091595,"type":"confidence_interval","upper":1459.7960466343015}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1457.3391897717306,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-29c268528b099a17007b","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-47eef41792434a5dec789e29","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":35.0,"rawValue":1506.8141348240842,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb60abb91d9cdc5d1cc20128e2160a0b8b7b780d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":429,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=428","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":429,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1506.8141348240842,"rating_lower":1498.7731035364384,"rating_upper":1514.8551661117301,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.83167442029677,"vote_count":6562.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1498.7731035364384,"type":"confidence_interval","upper":1514.8551661117301}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1506.8141348240842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f2e5b751dce32642b33f","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f35df1c9b13b313be505ed46","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":83.0,"rawValue":1452.4271145093594,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb60abb91d9cdc5d1cc20128e2160a0b8b7b780d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":841,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=840","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":841,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1452.4271145093594,"rating_lower":1448.1766367436214,"rating_upper":1456.6775922750971,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.703046962187101,"vote_count":26633.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.1766367436214,"type":"confidence_interval","upper":1456.6775922750971}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1452.4271145093594,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-883690ac90c1c35a22e6","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f392a193cb25322717236799","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12.0,"rawValue":1479.5500212955633,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d76930cec6e98e910944e8f18db865a10c156a2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":12,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=11","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":12,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1479.5500212955633,"rating_lower":1476.29092809673,"rating_upper":1482.8091144943965,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":2.765014275686347,"vote_count":99182.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1476.29092809673,"type":"confidence_interval","upper":1482.8091144943965}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1479.5500212955633,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f80f364152151fe7c781","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-59a93cd31c9d8da683ee4781","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":15.0,"rawValue":1531.4621627000524,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d76930cec6e98e910944e8f18db865a10c156a2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":409,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=408","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":409,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1531.4621627000524,"rating_lower":1522.247900534274,"rating_upper":1540.6764248658308,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.10166273352176,"vote_count":5615.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1522.247900534274,"type":"confidence_interval","upper":1540.6764248658308}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1531.4621627000524,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-30b82a98a287cc746753","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-54a58737f240c9d62b434c22","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":37.0,"rawValue":1483.3184123763037,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d76930cec6e98e910944e8f18db865a10c156a2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":795,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=794","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":795,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1483.3184123763037,"rating_lower":1478.298981270871,"rating_upper":1488.3378434817364,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.558625198963632,"vote_count":27416.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1478.298981270871,"type":"confidence_interval","upper":1488.3378434817364}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1483.3184123763037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-32ea4bce14033a978023","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7dc86d27ec9ed2ab603f2f21","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":9.0,"rawValue":1537.6543467895315,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c6d998ca46c2d3e0f9ccba5755708f33ea9e78","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":403,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=402","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":403,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1537.6543467895315,"rating_lower":1522.9933444567612,"rating_upper":1552.3153491223018,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":55.95400066338742,"vote_count":1910.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1522.9933444567612,"type":"confidence_interval","upper":1552.3153491223018}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1537.6543467895315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ed9c966771fbc8ae30d6","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b394a3c1fb05e562c2c3c894","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10.0,"rawValue":1482.6325665947309,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c6d998ca46c2d3e0f9ccba5755708f33ea9e78","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":10,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=9","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":10,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1482.6325665947309,"rating_lower":1478.1421250010803,"rating_upper":1487.1230081883814,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.24906465152303,"vote_count":30004.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1478.1421250010803,"type":"confidence_interval","upper":1487.1230081883814}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1482.6325665947309,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e0ac072be4d16b67e823","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ecf366af57b815b7c4f7723e","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22.0,"rawValue":1491.9294583881592,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c6d998ca46c2d3e0f9ccba5755708f33ea9e78","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":780,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=779","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":780,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1491.9294583881592,"rating_lower":1484.7834440115575,"rating_upper":1499.0754727647607,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.29326275619661,"vote_count":8771.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1484.7834440115575,"type":"confidence_interval","upper":1499.0754727647607}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1491.9294583881592,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d531d84c158edcb8eef4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5337677dc70f7f99d0020522","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":69.0,"rawValue":1474.565145064679,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ca30054e6d7400abc373211e9e862a576806657","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":463,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=462","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":463,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1474.565145064679,"rating_lower":1456.9812004006596,"rating_upper":1492.1490897286985,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":80.48898201944746,"vote_count":1261.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1456.9812004006596,"type":"confidence_interval","upper":1492.1490897286985}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1474.565145064679,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f58cf566f28f52d41960","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e066d0d3925d61eb84f1357a","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":73.0,"rawValue":1436.4808617125807,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ca30054e6d7400abc373211e9e862a576806657","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":73,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=72","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":73,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1436.4808617125807,"rating_lower":1430.978351950622,"rating_upper":1441.9833714745394,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.8818009234781705,"vote_count":17985.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.978351950622,"type":"confidence_interval","upper":1441.9833714745394}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1436.4808617125807,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ca46bcd796e98b60dbec","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-251b2ab681acc008f82b0372","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":85.0,"rawValue":1451.4549706628025,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ca30054e6d7400abc373211e9e862a576806657","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":843,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=842","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":843,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1451.4549706628025,"rating_lower":1442.5945278183929,"rating_upper":1460.3154135072123,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.436883763044925,"vote_count":5135.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1442.5945278183929,"type":"confidence_interval","upper":1460.3154135072123}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1451.4549706628025,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-187ef81985da66469f46","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bf4bc906a94c11d683ee76b4","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":18.0,"rawValue":1475.206155988065,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c6d998ca46c2d3e0f9ccba5755708f33ea9e78","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":18,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=17","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":18,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1475.206155988065,"rating_lower":1470.3624626744697,"rating_upper":1480.0498493016603,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.107410234304779,"vote_count":28343.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1470.3624626744697,"type":"confidence_interval","upper":1480.0498493016603}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1475.206155988065,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-11a446484b590a2f53df","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a11cb0134d5ac156eb69e07d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":23.0,"rawValue":1520.2097875893496,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c6d998ca46c2d3e0f9ccba5755708f33ea9e78","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":417,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=416","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":417,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1520.2097875893496,"rating_lower":1505.19687511168,"rating_upper":1535.2227000670196,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.67238244179579,"vote_count":1732.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1505.19687511168,"type":"confidence_interval","upper":1535.2227000670196}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1520.2097875893496,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-efd1dde7591be53218eb","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-01a1b0ba30809c8d881f3ad3","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":28.0,"rawValue":1487.6627859976325,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c6d998ca46c2d3e0f9ccba5755708f33ea9e78","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":786,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=785","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":786,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1487.6627859976325,"rating_lower":1480.3324879499073,"rating_upper":1494.9930840453576,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.987724970268825,"vote_count":8426.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1480.3324879499073,"type":"confidence_interval","upper":1494.9930840453576}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1487.6627859976325,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6b214ef24266246bf9cd","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a7830f3d08c7a8a1a278346a","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8.0,"rawValue":1541.8997963210757,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be71e377cb828219ec33f41b91f1e2cff714f259","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":402,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=401","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":402,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1541.8997963210757,"rating_lower":1523.3887364742607,"rating_upper":1560.4108561678904,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":89.20031494452974,"vote_count":1107.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1523.3887364742607,"type":"confidence_interval","upper":1560.4108561678904}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1541.8997963210757,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4221f1a8c686b18315bf","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e188a93a68e303167347a67c","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":16.0,"rawValue":1476.4838647133247,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be71e377cb828219ec33f41b91f1e2cff714f259","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":16,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=15","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":16,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1476.4838647133247,"rating_lower":1471.0556306240871,"rating_upper":1481.9120988025625,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.670451956644463,"vote_count":18018.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1471.0556306240871,"type":"confidence_interval","upper":1481.9120988025625}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1476.4838647133247,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f37710fdf47d10f9dbf7","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c943823991df56be9fb045a4","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":23.0,"rawValue":1491.6956027567426,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be71e377cb828219ec33f41b91f1e2cff714f259","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":781,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=780","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":781,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1491.6956027567426,"rating_lower":1483.0681984716914,"rating_upper":1500.3230070417937,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.375999632417916,"vote_count":5462.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1483.0681984716914,"type":"confidence_interval","upper":1500.3230070417937}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1491.6956027567426,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f078d2049dae2c14e4f9","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-44eeb791528061b6d175796c","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3.0,"rawValue":1562.1742490953136,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9173ff03fe1a6046411fb2c471db5dab304e0a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":397,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=396","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":397,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1562.1742490953136,"rating_lower":1530.8907235877646,"rating_upper":1593.4577746028626,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":254.76232177978574,"vote_count":385.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.8907235877646,"type":"confidence_interval","upper":1593.4577746028626}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1562.1742490953136,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-228f94b461edfb43e115","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2ae748d37d67edbee0966038","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":6.0,"rawValue":1490.2116753124994,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9173ff03fe1a6046411fb2c471db5dab304e0a9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":6,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=5","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":6,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1490.2116753124994,"rating_lower":1482.0807074514883,"rating_upper":1498.3426431735109,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.210294693423652,"vote_count":5718.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1482.0807074514883,"type":"confidence_interval","upper":1498.3426431735109}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1490.2116753124994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b43930230e234fb0a7f4","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cdc5b0bdf5c15ce1d8ca5e9c","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11.0,"rawValue":1503.5280062667107,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9173ff03fe1a6046411fb2c471db5dab304e0a9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":769,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=768","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":769,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1503.5280062667107,"rating_lower":1488.9598577399654,"rating_upper":1518.0961547934558,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":55.24748836405817,"vote_count":1756.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1488.9598577399654,"type":"confidence_interval","upper":1518.0961547934558}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1503.5280062667107,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-74bab3492ebc06fa1fa7","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7a8c7ee54ad9b2dea12575da","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":56.0,"rawValue":1485.77033925055,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9577e36d6b9e49ed6ad3d60a1c671f73b887acab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":450,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=449","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":450,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1485.77033925055,"rating_lower":1464.1248479853996,"rating_upper":1507.4158305157002,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":121.96598062843223,"vote_count":712.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1464.1248479853996,"type":"confidence_interval","upper":1507.4158305157002}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1485.77033925055,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d8cee9c329a556b3ee80","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5fed400a773ef90f7c49b98f","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":74.0,"rawValue":1455.7302170826042,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9577e36d6b9e49ed6ad3d60a1c671f73b887acab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":832,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=831","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":832,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1455.7302170826042,"rating_lower":1443.8726915291568,"rating_upper":1467.5877426360516,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.600916165815875,"vote_count":2428.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1443.8726915291568,"type":"confidence_interval","upper":1467.5877426360516}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1455.7302170826042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b989c5919e70245af8fe","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dcb858251fbad1bde1658385","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":75.0,"rawValue":1435.3354115275959,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9577e36d6b9e49ed6ad3d60a1c671f73b887acab","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":75,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=74","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":75,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1435.3354115275959,"rating_lower":1429.303276456019,"rating_upper":1441.3675465991728,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.472092561745482,"vote_count":12155.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1429.303276456019,"type":"confidence_interval","upper":1441.3675465991728}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1435.3354115275959,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3798f5bc1dbaaebdff73","canonicalModelId":"zai/glm-4.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9b18543e3f310f00b4ce05df","mappingCandidates":["zai/glm-4.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7 Flash","modelRef":"glm-4.7-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":141.0,"rawValue":1424.4551258868103,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02910d67999af8849cdccf458d89db9b9a4f7880","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":535,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=534","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":535,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1424.4551258868103,"rating_lower":1396.8590442678,"rating_upper":1452.0512075058205,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":198.2433643751701,"vote_count":440.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1396.8590442678,"type":"confidence_interval","upper":1452.0512075058205}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1424.4551258868103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-96f1ee6866e609bc38e1","canonicalModelId":"zai/glm-4.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0948ab4e0c8f561432646fd8","mappingCandidates":["zai/glm-4.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7 Flash","modelRef":"glm-4.7-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":169.0,"rawValue":1383.852604453216,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02910d67999af8849cdccf458d89db9b9a4f7880","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":927,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=926","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":927,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1383.852604453216,"rating_lower":1372.7398706185065,"rating_upper":1394.9653382879255,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.1473843779964,"vote_count":2694.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1372.7398706185065,"type":"confidence_interval","upper":1394.9653382879255}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1383.852604453216,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6da6a2e3f8f15f53fd8f","canonicalModelId":"zai/glm-4.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4063483a1830c1dfbbae32df","mappingCandidates":["zai/glm-4.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7 Flash","modelRef":"glm-4.7-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":184.0,"rawValue":1352.9259898135635,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02910d67999af8849cdccf458d89db9b9a4f7880","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":184,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=183","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":184,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1352.9259898135635,"rating_lower":1347.1849328146618,"rating_upper":1358.667046812465,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.580004889569823,"vote_count":11747.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1347.1849328146618,"type":"confidence_interval","upper":1358.667046812465}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1352.9259898135635,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4fcc51eee232bebc11f8","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3bd3bd9a99264121e065e395","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":30.0,"rawValue":1514.545958762338,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6af7fa0277a9883bef705ceaf1b5f3cbb8cb102e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":424,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=423","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":424,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1514.545958762338,"rating_lower":1498.6319076725413,"rating_upper":1530.4600098521346,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":65.92730363901099,"vote_count":1508.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1498.6319076725413,"type":"confidence_interval","upper":1530.4600098521346}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1514.545958762338,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-124d3fc83a1e29394190","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bd23eab86d49b31dacb8cbb8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":48.0,"rawValue":1445.2412180980982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6af7fa0277a9883bef705ceaf1b5f3cbb8cb102e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":48,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=47","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":48,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1445.2412180980982,"rating_lower":1440.9277515983458,"rating_upper":1449.5546845978504,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.843470700311699,"vote_count":27850.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1440.9277515983458,"type":"confidence_interval","upper":1449.5546845978504}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1445.2412180980982,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8d01425c6d42c93f7092","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d4a7be1bca5a8f2c6d03b256","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":69.0,"rawValue":1460.3721655134777,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6af7fa0277a9883bef705ceaf1b5f3cbb8cb102e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":827,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=826","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":827,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1460.3721655134777,"rating_lower":1453.0188879024056,"rating_upper":1467.7254431245497,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":14.075561954279324,"vote_count":7225.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1453.0188879024056,"type":"confidence_interval","upper":1467.7254431245497}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1460.3721655134777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47bef8222d5e043489d6","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-74dd6eb5d8ad21f557ccc9f8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":24.0,"rawValue":1520.10410144335,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2c57adb25415483fba6aefd6068737d103246ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":418,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=417","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":418,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1520.10410144335,"rating_lower":1507.4286033856788,"rating_upper":1532.7795995010213,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.82480107413853,"vote_count":2556.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1507.4286033856788,"type":"confidence_interval","upper":1532.7795995010213}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1520.10410144335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f9676d7dca45ee4a09a3","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ddf00a21f6e433fda4d5e6bb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":27.0,"rawValue":1488.1012064231313,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2c57adb25415483fba6aefd6068737d103246ba","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":785,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=784","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":785,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1488.1012064231313,"rating_lower":1481.732804870098,"rating_upper":1494.4696079761648,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.557587685750022,"vote_count":11848.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1481.732804870098,"type":"confidence_interval","upper":1494.4696079761648}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1488.1012064231313,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4a0bef17731963f5936a","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e51613ab4f5a9122e0a00b8f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":29.0,"rawValue":1464.107000003788,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2c57adb25415483fba6aefd6068737d103246ba","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":29,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=28","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":29,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1464.107000003788,"rating_lower":1460.048856607481,"rating_upper":1468.165143400095,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.287050465378726,"vote_count":42267.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.048856607481,"type":"confidence_interval","upper":1468.165143400095}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.107000003788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1421c2a6969a4434c753","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-304bb6e78630965076e8053a","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":27.0,"rawValue":1516.3461367724533,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d5df41a3380cdf332b08c8a0f4ca75297fb4063e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":421,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=420","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":421,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1516.3461367724533,"rating_lower":1502.0244696373586,"rating_upper":1530.667803907548,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":53.39381706332858,"vote_count":1886.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1502.0244696373586,"type":"confidence_interval","upper":1530.667803907548}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1516.3461367724533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-daec87f85d855a7b2dbf","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1bb3b53f8d78dfec7f0765e8","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":27.0,"rawValue":1465.3786574650403,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d5df41a3380cdf332b08c8a0f4ca75297fb4063e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":27,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=26","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":27,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1465.3786574650403,"rating_lower":1460.5391566691299,"rating_upper":1470.2181582609508,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.096842123479171,"vote_count":30637.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.5391566691299,"type":"confidence_interval","upper":1470.2181582609508}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1465.3786574650403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2d7c2d935e8e0c42237d","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dbc642135df06b48e9aad3bf","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":46.0,"rawValue":1475.473869793419,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d5df41a3380cdf332b08c8a0f4ca75297fb4063e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":804,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=803","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":804,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1475.473869793419,"rating_lower":1468.1940701426963,"rating_upper":1482.753669444142,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.79566602905527,"vote_count":8722.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1468.1940701426963,"type":"confidence_interval","upper":1482.753669444142}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1475.473869793419,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5bacdc17d952d32076ee","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8b0af01ee73e845754f6dc84","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":9.0,"rawValue":1504.0079599546075,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"148e058de0e4acf494eafde9e2bfb754e6c6e09f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":767,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=766","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":767,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1504.0079599546075,"rating_lower":1484.7632187611448,"rating_upper":1523.2527011480704,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":96.41130645686438,"vote_count":994.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1484.7632187611448,"type":"confidence_interval","upper":1523.2527011480704}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1504.0079599546075,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-162c4e5f9626f9371962","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f9467d9b5c03450f4b272eb4","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14.0,"rawValue":1531.996115754834,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"148e058de0e4acf494eafde9e2bfb754e6c6e09f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":408,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=407","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":408,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1531.996115754834,"rating_lower":1493.6593172143846,"rating_upper":1570.3329142952832,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":382.59166398285925,"vote_count":244.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1493.6593172143846,"type":"confidence_interval","upper":1570.3329142952832}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1531.996115754834,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-33181a94ae73f4275f49","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-eb5791f1973a201cd2a1169c","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":15.0,"rawValue":1476.929093287472,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"148e058de0e4acf494eafde9e2bfb754e6c6e09f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":15,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=14","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":15,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1476.929093287472,"rating_lower":1467.056715442164,"rating_upper":1486.8014711327796,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.371570767720776,"vote_count":3751.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1467.056715442164,"type":"confidence_interval","upper":1486.8014711327796}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1476.929093287472,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-352ff73df786ae88b96d","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c1ff19d552257ee3521cf9a8","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":50.0,"rawValue":1492.9304481628153,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a423189bbf189d52b4da737b9ddbccc1b71d16c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":444,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=443","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":444,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1492.9304481628153,"rating_lower":1468.8051772167528,"rating_upper":1517.0557191088776,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":151.51241374383744,"vote_count":599.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1468.8051772167528,"type":"confidence_interval","upper":1517.0557191088776}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1492.9304481628153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a03a651fd4d726a17d51","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-103558295d27ac7eb267f86d","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":58.0,"rawValue":1468.005293829955,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a423189bbf189d52b4da737b9ddbccc1b71d16c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":816,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=815","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":816,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1468.005293829955,"rating_lower":1456.1845753972077,"rating_upper":1479.826012262702,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.37404194301359,"vote_count":2620.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1456.1845753972077,"type":"confidence_interval","upper":1479.826012262702}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1468.005293829955,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a5b81ce74647f7a5682c","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-775ccc0bc001838bf3126be6","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":71.0,"rawValue":1436.7987275161317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a423189bbf189d52b4da737b9ddbccc1b71d16c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":71,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=70","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":71,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1436.7987275161317,"rating_lower":1430.1131529922625,"rating_upper":1443.4843020400012,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.635399154463647,"vote_count":9380.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.1131529922625,"type":"confidence_interval","upper":1443.4843020400012}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1436.7987275161317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ee4821dc2df6bad99ed0","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2d498bd59b01a6033c89d14f","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":109.0,"rawValue":1435.5240208841885,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5965216e4a2f5b30377431003b1f0e510fc9b97b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":867,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=866","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":867,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1435.5240208841885,"rating_lower":1427.6497827049304,"rating_upper":1443.3982590634469,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.140645988360237,"vote_count":6411.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1427.6497827049304,"type":"confidence_interval","upper":1443.3982590634469}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1435.5240208841885,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-692b30ae2460259da04e","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-69da1b7ae2bb86bc4f41450d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":131.0,"rawValue":1405.633007885001,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5965216e4a2f5b30377431003b1f0e510fc9b97b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":131,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=130","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":131,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1405.633007885001,"rating_lower":1401.223269385369,"rating_upper":1410.042746384633,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.062085666617317,"vote_count":32081.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1401.223269385369,"type":"confidence_interval","upper":1410.042746384633}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1405.633007885001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-86e421d345b56dc59509","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1c9d362f062c5db37ac39993","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":143.0,"rawValue":1423.687412475318,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5965216e4a2f5b30377431003b1f0e510fc9b97b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":537,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=536","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":537,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1423.687412475318,"rating_lower":1408.3243150384167,"rating_upper":1439.0505099122192,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":61.44144031539489,"vote_count":1618.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1408.3243150384167,"type":"confidence_interval","upper":1439.0505099122192}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1423.687412475318,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4986120771a647cd4697","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a3570b0debd3c747515cec65","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":147.0,"rawValue":1406.7245246472921,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f1d098bda5ccc2bae249e655aaf2ad79c998a04e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":905,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=904","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":905,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1406.7245246472921,"rating_lower":1398.4567142332166,"rating_upper":1414.9923350613676,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.794460967498626,"vote_count":5539.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1398.4567142332166,"type":"confidence_interval","upper":1414.9923350613676}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1406.7245246472921,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c5011c6766ec3003eb93","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2d027360a2881ed477d77911","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":161.0,"rawValue":1373.4392475339996,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f1d098bda5ccc2bae249e655aaf2ad79c998a04e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":161,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=160","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":161,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1373.4392475339996,"rating_lower":1368.907006582511,"rating_upper":1377.9714884854882,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.347241504111302,"vote_count":27140.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1368.907006582511,"type":"confidence_interval","upper":1377.9714884854882}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1373.4392475339996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-157144b3754f41717eea","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f82cde7143b716ddb1ea69df","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":166.0,"rawValue":1386.3329871664873,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f1d098bda5ccc2bae249e655aaf2ad79c998a04e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":560,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=559","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":560,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1386.3329871664873,"rating_lower":1369.8296004940635,"rating_upper":1402.8363738389116,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":70.90060947479643,"vote_count":1361.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1369.8296004940635,"type":"confidence_interval","upper":1402.8363738389116}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1386.3329871664873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8991e9e565fc685ce044","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-279641ec268c28b2569053fc","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":191.0,"rawValue":1349.6979522075699,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffb90d7a7951ce61f7794aca0dead6d8b39fb1e3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":585,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=584","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":585,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1349.6979522075699,"rating_lower":1320.08884396224,"rating_upper":1379.3070604528998,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":228.22040584187548,"vote_count":383.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1320.08884396224,"type":"confidence_interval","upper":1379.3070604528998}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1349.6979522075699,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3dbdbddac162050244f7","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2c45441a73d732fb9e72a31c","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":203.0,"rawValue":1352.4736004900342,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffb90d7a7951ce61f7794aca0dead6d8b39fb1e3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":961,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=960","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":961,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1352.4736004900342,"rating_lower":1337.9757694147559,"rating_upper":1366.9714315653125,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":54.71543902931328,"vote_count":1684.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1337.9757694147559,"type":"confidence_interval","upper":1366.9714315653125}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1352.4736004900342,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-97b92d20e18d354171ab","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0a90630ce620b718ee0f4c2b","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":212.0,"rawValue":1320.2673210882358,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffb90d7a7951ce61f7794aca0dead6d8b39fb1e3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":212,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=211","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":212,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1320.2673210882358,"rating_lower":1313.4797642396397,"rating_upper":1327.0548779368319,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.993081306699434,"vote_count":8292.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1313.4797642396397,"type":"confidence_interval","upper":1327.0548779368319}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1320.2673210882358,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9135a741ae487c51f13b","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dba3ae933c6fe4f76855c264","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":67.0,"rawValue":1474.9413099648884,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833242f1e654f31c549aaf1ab4ed5f2949356f15","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":461,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=460","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":461,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1474.9413099648884,"rating_lower":1461.7663346187296,"rating_upper":1488.1162853110475,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":45.185952387882146,"vote_count":2083.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1461.7663346187296,"type":"confidence_interval","upper":1488.1162853110475}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1474.9413099648884,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9589067ed0b0e40eaf44","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d18043e5964127d79b38bd76","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":95.0,"rawValue":1422.2190385742006,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833242f1e654f31c549aaf1ab4ed5f2949356f15","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":95,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=94","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":95,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1422.2190385742006,"rating_lower":1418.6181507699587,"rating_upper":1425.8199263784425,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.375382526265093,"vote_count":43653.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1418.6181507699587,"type":"confidence_interval","upper":1425.8199263784425}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1422.2190385742006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fec356391dbb3b2d10bc","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b0cf7ee9862679ed54a188f7","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":107.0,"rawValue":1436.3313060625342,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833242f1e654f31c549aaf1ab4ed5f2949356f15","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":865,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=864","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":865,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1436.3313060625342,"rating_lower":1429.674410719313,"rating_upper":1442.9882014057553,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.535788271860556,"vote_count":9153.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1429.674410719313,"type":"confidence_interval","upper":1442.9882014057553}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1436.3313060625342,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-881999bf955a16597f92","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f83b7e31be61501fd2ddbd29","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":47.0,"rawValue":1495.4011508648664,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833242f1e654f31c549aaf1ab4ed5f2949356f15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":441,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=440","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":441,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1495.4011508648664,"rating_lower":1482.0465197712786,"rating_upper":1508.7557819584545,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":46.426677981048535,"vote_count":2069.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1482.0465197712786,"type":"confidence_interval","upper":1508.7557819584545}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1495.4011508648664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-69230f643c3a800c0006","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b0a3e236ae308f62b85665e8","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":59.0,"rawValue":1441.3712036369134,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833242f1e654f31c549aaf1ab4ed5f2949356f15","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":59,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=58","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":59,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1441.3712036369134,"rating_lower":1437.6461428603334,"rating_upper":1445.0962644134934,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.6121896490114356,"vote_count":40945.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1437.6461428603334,"type":"confidence_interval","upper":1445.0962644134934}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1441.3712036369134,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6f49be1713fc65a75277","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7c36f015c51f948b52741790","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":84.0,"rawValue":1452.413214579112,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833242f1e654f31c549aaf1ab4ed5f2949356f15","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":842,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=841","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":842,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1452.413214579112,"rating_lower":1445.436650638128,"rating_upper":1459.3897785200961,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.670302271740857,"vote_count":8232.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1445.436650638128,"type":"confidence_interval","upper":1459.3897785200961}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1452.413214579112,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1d03282499ee0d6a952e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2822f01ac746e1e15cbba927","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":116.0,"rawValue":1431.7860968982502,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a60c2838df2ad462aa7d6fd16746fe868cba2f74","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":874,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=873","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":874,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1431.7860968982502,"rating_lower":1426.5388988466837,"rating_upper":1437.0332949498168,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.1673519559915935,"vote_count":20858.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1426.5388988466837,"type":"confidence_interval","upper":1437.0332949498168}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1431.7860968982502,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-746a8ba0e6138c5127cf","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ca9fe72a445ff817d5559cd8","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":118.0,"rawValue":1449.207461053533,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a60c2838df2ad462aa7d6fd16746fe868cba2f74","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":512,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=511","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":512,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1449.207461053533,"rating_lower":1438.9982257508425,"rating_upper":1459.4166963562236,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.132527076489676,"vote_count":4020.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1438.9982257508425,"type":"confidence_interval","upper":1459.4166963562236}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1449.207461053533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3257e5daac74e5253427","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c345d04a0f1891bec2ad9c2f","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":121.0,"rawValue":1411.9324804584382,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a60c2838df2ad462aa7d6fd16746fe868cba2f74","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":121,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=120","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":121,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1411.9324804584382,"rating_lower":1408.7081138567116,"rating_upper":1415.1568470601646,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":2.70640412083102,"vote_count":79397.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1408.7081138567116,"type":"confidence_interval","upper":1415.1568470601646}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1411.9324804584382,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c8fd85b838099ab1f784","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6349aea34a82d615e06c77ed","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":97.0,"rawValue":1442.4392042103425,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a60c2838df2ad462aa7d6fd16746fe868cba2f74","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":855,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=854","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":855,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1442.4392042103425,"rating_lower":1436.3029270379836,"rating_upper":1448.5754813827014,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.801978699646211,"vote_count":11742.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1436.3029270379836,"type":"confidence_interval","upper":1448.5754813827014}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1442.4392042103425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-08e6fdd8a2d63a7a0dae","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e306fda394d1f9ebec961faf","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":114.0,"rawValue":1450.6337964584836,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a60c2838df2ad462aa7d6fd16746fe868cba2f74","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":508,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=507","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":508,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1450.6337964584836,"rating_lower":1438.3792642508856,"rating_upper":1462.8883286660816,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.09284639941125,"vote_count":2558.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1438.3792642508856,"type":"confidence_interval","upper":1462.8883286660816}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1450.6337964584836,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-eda974bf964bbe657566","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-12d4c284738adaba05fcecc4","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":114.0,"rawValue":1416.4915639243197,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a60c2838df2ad462aa7d6fd16746fe868cba2f74","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":114,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=113","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":114,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1416.4915639243197,"rating_lower":1412.9057657146593,"rating_upper":1420.0773621339802,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.3471525794152908,"vote_count":47920.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1412.9057657146593,"type":"confidence_interval","upper":1420.0773621339802}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1416.4915639243197,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-eb939ef5ac4a7588f18d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f0d6554c940c1738aff26dd9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":33.0,"rawValue":1509.7354893174718,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07d8366f45035b5cf72ec300ae95987590dd9e57","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":427,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=426","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":427,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1509.7354893174718,"rating_lower":1498.6969515511762,"rating_upper":1520.7740270837671,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.7195424200622,"vote_count":3555.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1498.6969515511762,"type":"confidence_interval","upper":1520.7740270837671}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1509.7354893174718,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d619c0c73ca4dded163b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3b376d053d6a02967cda0f0d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":36.0,"rawValue":1452.5898889549094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07d8366f45035b5cf72ec300ae95987590dd9e57","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":36,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=35","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":36,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1452.5898889549094,"rating_lower":1448.8111056088674,"rating_upper":1456.3686723009514,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.717130455597323,"vote_count":63659.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.8111056088674,"type":"confidence_interval","upper":1456.3686723009514}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1452.5898889549094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-acb1aa9266dbcc81c0b4","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5f1e69a918f0d82b7c22deb4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":42.0,"rawValue":1480.1373965230498,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07d8366f45035b5cf72ec300ae95987590dd9e57","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":800,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=799","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":800,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1480.1373965230498,"rating_lower":1474.2384757262816,"rating_upper":1486.036317319818,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.058346891313478,"vote_count":17712.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1474.2384757262816,"type":"confidence_interval","upper":1486.036317319818}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1480.1373965230498,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e3bcc5f269653fbfc559","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4b9b551f0147b600d21ba83d","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":17.0,"rawValue":1495.808303356952,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07d8366f45035b5cf72ec300ae95987590dd9e57","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":775,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=774","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":775,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1495.808303356952,"rating_lower":1489.8494254968816,"rating_upper":1501.767181217022,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.243422097862977,"vote_count":16467.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1489.8494254968816,"type":"confidence_interval","upper":1501.767181217022}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1495.808303356952,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-30a3d93ba5e10f3873b5","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-527017855db531446e9c3460","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":23.0,"rawValue":1469.5687129457133,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07d8366f45035b5cf72ec300ae95987590dd9e57","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":23,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=22","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":23,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1469.5687129457133,"rating_lower":1465.7118379367255,"rating_upper":1473.4255879547009,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.8723530641064987,"vote_count":60708.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1465.7118379367255,"type":"confidence_interval","upper":1473.4255879547009}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1469.5687129457133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0c66f4600993bd2bbac0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7285aefab3599a081162bdba","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":25.0,"rawValue":1518.363983589094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07d8366f45035b5cf72ec300ae95987590dd9e57","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":419,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=418","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":419,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1518.363983589094,"rating_lower":1507.071144309347,"rating_upper":1529.6568228688411,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":33.197861789172244,"vote_count":3347.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1507.071144309347,"type":"confidence_interval","upper":1529.6568228688411}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1518.363983589094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-984d0c980e8799340449","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3e3d85510d16d10211955eb1","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":102.0,"rawValue":1439.085991867134,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5081f7eb998c19993da8f1eb0c6b2041ca1408b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":860,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=859","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":860,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1439.085991867134,"rating_lower":1433.0894237631196,"rating_upper":1445.0825599711484,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.360722242386556,"vote_count":16230.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1433.0894237631196,"type":"confidence_interval","upper":1445.0825599711484}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1439.085991867134,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6d22586d786cc631772a","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-07f7de66eb9329380124ba86","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":116.0,"rawValue":1450.0267388282168,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5081f7eb998c19993da8f1eb0c6b2041ca1408b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":510,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=509","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":510,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1450.0267388282168,"rating_lower":1438.5784855935167,"rating_upper":1461.474992062917,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.117898497254274,"vote_count":3222.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1438.5784855935167,"type":"confidence_interval","upper":1461.474992062917}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1450.0267388282168,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4635f8ac66d5afdb0b0f","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-43f867c038f73c827e6c08bb","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":120.0,"rawValue":1412.1044098861782,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5081f7eb998c19993da8f1eb0c6b2041ca1408b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":120,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=119","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":120,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1412.1044098861782,"rating_lower":1408.2778339557046,"rating_upper":1415.9309858166516,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.8117507007490374,"vote_count":59502.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1408.2778339557046,"type":"confidence_interval","upper":1415.9309858166516}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1412.1044098861782,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d322b7089396a81a6dd0","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-58873118bc0643218cdf6ee7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":151.0,"rawValue":1404.044665924795,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a920774bc51afa4cb17fece5c5abf817f7267fff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":909,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=908","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":909,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1404.044665924795,"rating_lower":1398.0084961221398,"rating_upper":1410.0808357274504,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.484768049629631,"vote_count":16573.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1398.0084961221398,"type":"confidence_interval","upper":1410.0808357274504}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1404.044665924795,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e1f9da22c33991efee1c","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f8e079db6ac622c9cb5f3d46","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":162.0,"rawValue":1390.7940153666893,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a920774bc51afa4cb17fece5c5abf817f7267fff","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":556,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=555","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":556,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1390.7940153666893,"rating_lower":1379.0729584963121,"rating_upper":1402.5150722370665,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.76328175601609,"vote_count":3144.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1379.0729584963121,"type":"confidence_interval","upper":1402.5150722370665}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1390.7940153666893,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bcf144900385089f97a0","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c5c44201f69e931af213227a","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":162.0,"rawValue":1372.7685819566527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a920774bc51afa4cb17fece5c5abf817f7267fff","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":162,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=161","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":162,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1372.7685819566527,"rating_lower":1368.9080788883325,"rating_upper":1376.629085024973,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.8796417288723744,"vote_count":58551.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1368.9080788883325,"type":"confidence_interval","upper":1376.629085024973}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1372.7685819566527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4cd049da6bf1449bba86","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7365d0cc744afee5b9c45231","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":10.0,"rawValue":1537.421654292945,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":404,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=403","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":404,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1537.421654292945,"rating_lower":1525.8332928696698,"rating_upper":1549.0100157162203,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.9581049139529,"vote_count":3249.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1525.8332928696698,"type":"confidence_interval","upper":1549.0100157162203}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1537.421654292945,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-95ce1c812c7600ca7be5","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a96192f3263ae998d06b0704","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":26.0,"rawValue":1465.8000166386585,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":26,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=25","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":26,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1465.8000166386585,"rating_lower":1461.9075016551055,"rating_upper":1469.6925316222114,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.9442497250162734,"vote_count":60896.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1461.9075016551055,"type":"confidence_interval","upper":1469.6925316222114}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1465.8000166386585,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cae870b9fb9ca98d4af1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8020cb7229b3332aa722bf29","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":41.0,"rawValue":1481.642180792169,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":799,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=798","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":799,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1481.642180792169,"rating_lower":1475.7098534547479,"rating_upper":1487.5745081295902,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.161235166372244,"vote_count":17279.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1475.7098534547479,"type":"confidence_interval","upper":1487.5745081295902}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1481.642180792169,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-11a60295358489565db6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e9d2dd548bc8e9b49d0965b6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18.0,"rawValue":1494.471152235676,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":776,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=775","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":776,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1494.471152235676,"rating_lower":1488.4570921212337,"rating_upper":1500.4852123501182,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.415412411889198,"vote_count":16589.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1488.4570921212337,"type":"confidence_interval","upper":1500.4852123501182}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1494.471152235676,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-94bf25666c2167c66847","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8183d6a83c0323728b11fb5b","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":22.0,"rawValue":1525.3095553781413,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":416,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=415","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":416,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1525.3095553781413,"rating_lower":1514.122246015761,"rating_upper":1536.4968647405217,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.580302591136,"vote_count":3379.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1514.122246015761,"type":"confidence_interval","upper":1536.4968647405217}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1525.3095553781413,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c02c5401c0ecedfdd374","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3b88a37c10fe03d86c5944ae","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":22.0,"rawValue":1470.8780077273245,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":22,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=21","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":22,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1470.8780077273245,"rating_lower":1466.9596472041692,"rating_upper":1474.7963682504799,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.9968017115560697,"vote_count":59545.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1466.9596472041692,"type":"confidence_interval","upper":1474.7963682504799}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1470.8780077273245,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-51fbfb07d0a6db20583d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-784d0ed138be1e70dc43b425","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":99.0,"rawValue":1458.0333757430712,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":493,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=492","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":493,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1458.0333757430712,"rating_lower":1441.6582062330558,"rating_upper":1474.408545253087,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":69.80321513202907,"vote_count":1485.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1441.6582062330558,"type":"confidence_interval","upper":1474.408545253087}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1458.0333757430712,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-415860f953dd0d7c09aa","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-15aad823bca8b5f0dcc46d47","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":103.0,"rawValue":1419.5386838826466,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":103,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=102","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":103,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1419.5386838826466,"rating_lower":1414.4939853030226,"rating_upper":1424.5833824622707,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.6248227423826815,"vote_count":25863.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1414.4939853030226,"type":"confidence_interval","upper":1424.5833824622707}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1419.5386838826466,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf1cf784ac972cd8bada","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d9b23fffb3bc508cc9fa8e7f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":111.0,"rawValue":1434.207722401864,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a62203b9ca43dcb85ebf9b3767f4d0ffbcac9030","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":869,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=868","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":869,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1434.207722401864,"rating_lower":1426.4529834709733,"rating_upper":1441.9624613327549,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.654463237329672,"vote_count":7512.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1426.4529834709733,"type":"confidence_interval","upper":1441.9624613327549}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1434.207722401864,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a2052a6975af131762db","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6c43980bd1293502685ac786","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":65.0,"rawValue":1463.4279627739224,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8aff69f6739d10b02589c48ce42687561981fef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":823,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=822","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":823,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1463.4279627739224,"rating_lower":1455.0602449236485,"rating_upper":1471.7956806241964,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.22711248252872,"vote_count":5888.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1455.0602449236485,"type":"confidence_interval","upper":1471.7956806241964}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1463.4279627739224,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-93b1febfbc1595460a8b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b4b9c2d0558aef514cc96e5c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":82.0,"rawValue":1467.224316956766,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8aff69f6739d10b02589c48ce42687561981fef","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":476,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=475","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":476,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1467.224316956766,"rating_lower":1450.8492635581551,"rating_upper":1483.5993703553768,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":69.80222522830661,"vote_count":1471.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1450.8492635581551,"type":"confidence_interval","upper":1483.5993703553768}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1467.224316956766,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-629b20305b293049ad6a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4e84c51d02f6d6552342f9da","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":83.0,"rawValue":1428.5368364471099,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8aff69f6739d10b02589c48ce42687561981fef","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":83,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=82","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":83,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1428.5368364471099,"rating_lower":1423.2980632397212,"rating_upper":1433.7756096544983,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.14435478797946,"vote_count":20693.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1423.2980632397212,"type":"confidence_interval","upper":1433.7756096544983}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1428.5368364471099,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3316255a2e2bb63465a7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fdfc57aea97e1ef42bb60e19","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13.0,"rawValue":1532.2918933515473,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae014511ca843eaff1c3b36b8601bbabbbbac6ef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":407,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=406","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":407,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1532.2918933515473,"rating_lower":1515.041253323049,"rating_upper":1549.542533380046,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":77.46655509873838,"vote_count":1320.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1515.041253323049,"type":"confidence_interval","upper":1549.542533380046}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1532.2918933515473,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-09ef3c8c561cdf831746","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0d387c6a952a402ef0a2fc24","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":21.0,"rawValue":1492.2691050475642,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae014511ca843eaff1c3b36b8601bbabbbbac6ef","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":779,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=778","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":779,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1492.2691050475642,"rating_lower":1483.6689695023,"rating_upper":1500.8692405928286,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.253709293584,"vote_count":5490.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1483.6689695023,"type":"confidence_interval","upper":1500.8692405928286}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1492.2691050475642,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a1a4eda39c945b3de669","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e80ac457acfa32fa40ae33e7","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":35.0,"rawValue":1454.1709198249014,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae014511ca843eaff1c3b36b8601bbabbbbac6ef","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":35,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=34","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":35,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1454.1709198249014,"rating_lower":1448.7758009463632,"rating_upper":1459.5660387034395,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.577149482055138,"vote_count":19541.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.7758009463632,"type":"confidence_interval","upper":1459.5660387034395}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1454.1709198249014,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-87cb23e33ee311c9ae8d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-223dbbe1cc840297b5d2e31d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":28.0,"rawValue":1515.565205789121,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"581ec3a7deab89f9945a3efdce12c50d84e9349d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":422,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=421","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":422,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1515.565205789121,"rating_lower":1499.0617653387667,"rating_upper":1532.068646239475,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":70.90107154895107,"vote_count":1398.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1499.0617653387667,"type":"confidence_interval","upper":1532.068646239475}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1515.565205789121,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b7c616fb51db3d193370","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cdd3adda03d2748bdb534616","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":33.0,"rawValue":1484.9891712747985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"581ec3a7deab89f9945a3efdce12c50d84e9349d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":791,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=790","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":791,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1484.9891712747985,"rating_lower":1476.5209429461775,"rating_upper":1493.4573996034196,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.667619353187256,"vote_count":5622.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1476.5209429461775,"type":"confidence_interval","upper":1493.4573996034196}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1484.9891712747985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-25edadbf3ceffab679b1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1e0a12777bf5d864d3da0d57","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":47.0,"rawValue":1446.0343010991212,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"581ec3a7deab89f9945a3efdce12c50d84e9349d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":47,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=46","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":47,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1446.0343010991212,"rating_lower":1440.747456600134,"rating_upper":1451.3211455981084,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.2760703839643925,"vote_count":20216.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1440.747456600134,"type":"confidence_interval","upper":1451.3211455981084}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1446.0343010991212,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf674c6147e3e7b68bea","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8aaef5fc5d193c39594de755","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-reasoning","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":127.0,"rawValue":1435.5106103793898,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7acfef7887608c4968d1cec3c84ec5dc59283726","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":521,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=520","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":521,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1435.5106103793898,"rating_lower":1413.3618044932819,"rating_upper":1457.6594162654978,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":127.70398566757164,"vote_count":721.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1413.3618044932819,"type":"confidence_interval","upper":1457.6594162654978}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1435.5106103793898,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e75760314fed3310ad39","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-05bb671297e0db857d6b082c","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-reasoning","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":132.0,"rawValue":1418.1419413595122,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7acfef7887608c4968d1cec3c84ec5dc59283726","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":890,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=889","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":890,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1418.1419413595122,"rating_lower":1408.831705990211,"rating_upper":1427.4521767288134,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.564470082260062,"vote_count":3973.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1408.831705990211,"type":"confidence_interval","upper":1427.4521767288134}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1418.1419413595122,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0c0da17f4f84c26b9c8b","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-006fc75e82b164b23126af1c","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-reasoning","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":139.0,"rawValue":1397.9462902761322,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7acfef7887608c4968d1cec3c84ec5dc59283726","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":139,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=138","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":139,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1397.9462902761322,"rating_lower":1392.9586696714762,"rating_upper":1402.9339108807885,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.475758418124843,"vote_count":18811.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1392.9586696714762,"type":"confidence_interval","upper":1402.9339108807885}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1397.9462902761322,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9adec681dde00e7f2caa","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f20ba3dfce7e751fb02a04f6","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37.0,"rawValue":1452.2990764053388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dd842833997057df0f8c243f4140e9c65928245","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":37,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=36","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":37,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1452.2990764053388,"rating_lower":1447.146387766874,"rating_upper":1457.4517650438033,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.91148895360683,"vote_count":22030.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.146387766874,"type":"confidence_interval","upper":1457.4517650438033}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1452.2990764053388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9c0c58a573f48664712f","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-79c527640a9c165ce774b2a8","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38.0,"rawValue":1483.0229340466567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dd842833997057df0f8c243f4140e9c65928245","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":796,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=795","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":796,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1483.0229340466567,"rating_lower":1474.8321950565912,"rating_upper":1491.213673036722,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.464252080998307,"vote_count":6313.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1474.8321950565912,"type":"confidence_interval","upper":1491.213673036722}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1483.0229340466567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a49dc1572bf6611983be","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fbf2d320cd7d1eff898035bf","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":43.0,"rawValue":1499.7382999269346,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dd842833997057df0f8c243f4140e9c65928245","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":437,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=436","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":437,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1499.7382999269346,"rating_lower":1483.3426549889832,"rating_upper":1516.133944864886,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":69.97788743256525,"vote_count":1462.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1483.3426549889832,"type":"confidence_interval","upper":1516.133944864886}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1499.7382999269346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8edf9f7a2eec8e5fd097","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-efd85b6d9f78f2cbb8383eca","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17.0,"rawValue":1529.4649127956318,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eab8e74cc98bff32296c4c2871df1a00cc36ba32","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":411,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=410","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":411,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1529.4649127956318,"rating_lower":1495.227755431357,"rating_upper":1563.7020701599063,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":305.14005202177094,"vote_count":290.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1495.227755431357,"type":"confidence_interval","upper":1563.7020701599063}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1529.4649127956318,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7e7a4886fe8a534532a2","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d62f0ff0dba278ffd9464148","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":48.0,"rawValue":1474.963430362601,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eab8e74cc98bff32296c4c2871df1a00cc36ba32","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":806,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=805","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":806,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1474.963430362601,"rating_lower":1454.545081816366,"rating_upper":1495.3817789088362,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":108.5288107501325,"vote_count":845.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1454.545081816366,"type":"confidence_interval","upper":1495.3817789088362}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1474.963430362601,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b8e11fec745b689ea732","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8001312bb8c81d7750ffce9c","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":54.0,"rawValue":1443.7345059556712,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eab8e74cc98bff32296c4c2871df1a00cc36ba32","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":54,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=53","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":54,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1443.7345059556712,"rating_lower":1433.6623242044946,"rating_upper":1453.806687706848,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":26.40893732407785,"vote_count":3473.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1433.6623242044946,"type":"confidence_interval","upper":1453.806687706848}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1443.7345059556712,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-faa238d09de6694b2134","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fb683844046426bb5815dc63","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35.0,"rawValue":1484.2317973593354,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b97fad605bf8afe5d66f61fa075f0b9bf24db56b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":793,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=792","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":793,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1484.2317973593354,"rating_lower":1470.5823099896018,"rating_upper":1497.881284729069,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.49941497559839,"vote_count":1810.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1470.5823099896018,"type":"confidence_interval","upper":1497.881284729069}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1484.2317973593354,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1d3f2cb98a69589ee0b5","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-accf314ad770bd8b30bfc029","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":94.0,"rawValue":1460.7031659670874,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b97fad605bf8afe5d66f61fa075f0b9bf24db56b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":488,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=487","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":488,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1460.7031659670874,"rating_lower":1423.608950856108,"rating_upper":1497.797381078067,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":358.19225427776,"vote_count":242.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1423.608950856108,"type":"confidence_interval","upper":1497.797381078067}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1460.7031659670874,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-86a734daa21251dfd416","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2c19322d53b5d62aa9580f32","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":99.0,"rawValue":1420.3454023693812,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b97fad605bf8afe5d66f61fa075f0b9bf24db56b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":99,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=98","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":99,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1420.3454023693812,"rating_lower":1413.8095707574032,"rating_upper":1426.881233981359,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.120018944368379,"vote_count":8186.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1413.8095707574032,"type":"confidence_interval","upper":1426.881233981359}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1420.3454023693812,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9d04e008f4f62b295549","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4a0d11c8f5a766da1aabe8a6","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42.0,"rawValue":1500.265497589836,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b97fad605bf8afe5d66f61fa075f0b9bf24db56b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":436,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=435","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":436,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1500.265497589836,"rating_lower":1489.523592656043,"rating_upper":1511.007402523629,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":30.037682815977053,"vote_count":3515.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1489.523592656043,"type":"confidence_interval","upper":1511.007402523629}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1500.265497589836,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2a54c0f7e5d61fb9a500","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-334ec56652cca04554d1dcee","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":49.0,"rawValue":1445.2327043266241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b97fad605bf8afe5d66f61fa075f0b9bf24db56b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":49,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=48","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":49,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1445.2327043266241,"rating_lower":1441.891872391599,"rating_upper":1448.5735362616492,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":2.9054477840445014,"vote_count":70943.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1441.891872391599,"type":"confidence_interval","upper":1448.5735362616492}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1445.2327043266241,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-23914076115436443988","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9770bd46eeb752dd545daf5a","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":51.0,"rawValue":1473.2918632879823,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b97fad605bf8afe5d66f61fa075f0b9bf24db56b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":809,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=808","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":809,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1473.2918632879823,"rating_lower":1468.0139667802791,"rating_upper":1478.5697597956857,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.2514617092766365,"vote_count":19675.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1468.0139667802791,"type":"confidence_interval","upper":1478.5697597956857}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1473.2918632879823,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-48781ccea969a0fa8f9d","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ae621c9909ee3492a97363ec","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":16.0,"rawValue":1529.8696294273634,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc89ade0d7029c4473f123dd7045cc1eeed15982","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":410,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=409","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":410,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1529.8696294273634,"rating_lower":1515.400785039445,"rating_upper":1544.3384738152818,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":54.49686374198005,"vote_count":1901.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1515.400785039445,"type":"confidence_interval","upper":1544.3384738152818}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1529.8696294273634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8b3d4957947e09cff75e","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-16d1fe61507ad66a39500bcf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":30.0,"rawValue":1486.734867015527,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc89ade0d7029c4473f123dd7045cc1eeed15982","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":788,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=787","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":788,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1486.734867015527,"rating_lower":1479.7366979185447,"rating_upper":1493.7330361125094,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.748899050051103,"vote_count":10418.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1479.7366979185447,"type":"confidence_interval","upper":1493.7330361125094}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1486.734867015527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-20cc7e92bd228b1d050e","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7b133655be3c95e292075139","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":34.0,"rawValue":1455.0168217501182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc89ade0d7029c4473f123dd7045cc1eeed15982","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":34,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=33","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":34,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1455.0168217501182,"rating_lower":1450.490780687369,"rating_upper":1459.5428628128675,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.332621969377401,"vote_count":37543.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1450.490780687369,"type":"confidence_interval","upper":1459.5428628128675}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1455.0168217501182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e2075b572cde09d050a6","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6e249e7512cce04703715fa6","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":8.0,"rawValue":1511.541768583301,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fc901746c197e38881f74aee0e3f60b129453df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":766,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=765","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":766,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1511.541768583301,"rating_lower":1501.7325573291444,"rating_upper":1521.3509798374578,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.04793879614952,"vote_count":3966.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.7325573291444,"type":"confidence_interval","upper":1521.3509798374578}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1511.541768583301,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7cc422a1dc392cc66079","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d6953c5fe748882bd871729c","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":17.0,"rawValue":1476.2899572288309,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fc901746c197e38881f74aee0e3f60b129453df","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":17,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=16","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":17,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1476.2899572288309,"rating_lower":1470.4979946528929,"rating_upper":1482.0819198047689,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.73283615597999,"vote_count":15054.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1470.4979946528929,"type":"confidence_interval","upper":1482.0819198047689}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1476.2899572288309,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7bdc959941c4ff7f1ceb","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-265fa8abcd885feaee917c54","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":29.0,"rawValue":1515.482090092387,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fc901746c197e38881f74aee0e3f60b129453df","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":423,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=422","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":423,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1515.482090092387,"rating_lower":1496.1673214635764,"rating_upper":1534.7968587211974,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":97.11422264239624,"vote_count":986.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1496.1673214635764,"type":"confidence_interval","upper":1534.7968587211974}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1515.482090092387,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6397a1975ccd4a23cded","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f95b1a65662b1ee212ee93d9","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":87.0,"rawValue":1466.0034476538067,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a175a8ab65c8640f69817e77f12994d0ffa943f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":481,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=480","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":481,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1466.0034476538067,"rating_lower":1453.4760754078332,"rating_upper":1478.5308198997805,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.85298390907528,"vote_count":2379.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1453.4760754078332,"type":"confidence_interval","upper":1478.5308198997805}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1466.0034476538067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-53e5d8d5003993bfdc97","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4338d83a2ee07ff441b736b8","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":99.0,"rawValue":1440.9802098647078,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a175a8ab65c8640f69817e77f12994d0ffa943f4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":857,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=856","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":857,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1440.9802098647078,"rating_lower":1435.0063576362618,"rating_upper":1446.9540620931539,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.289937003896293,"vote_count":11991.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1435.0063576362618,"type":"confidence_interval","upper":1446.9540620931539}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1440.9802098647078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1ec6038aa577693aa2ca","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bbbbf877f6af079c3c67fa23","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":123.0,"rawValue":1411.1437952958465,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a175a8ab65c8640f69817e77f12994d0ffa943f4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":123,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=122","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":123,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1411.1437952958465,"rating_lower":1407.6452654623852,"rating_upper":1414.6423251293081,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.1862142917384633,"vote_count":46603.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1407.6452654623852,"type":"confidence_interval","upper":1414.6423251293081}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1411.1437952958465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-852ca55376217825a480","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-27807917fec3611702b025ec","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":129.0,"rawValue":1432.6867312156085,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a175a8ab65c8640f69817e77f12994d0ffa943f4","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":523,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=522","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":523,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1432.6867312156085,"rating_lower":1405.8686515889358,"rating_upper":1459.5048108422811,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":187.2229870038281,"vote_count":450.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1405.8686515889358,"type":"confidence_interval","upper":1459.5048108422811}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1432.6867312156085,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4de9f7b89588f509102a","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b1fbd6d98b97d1cd014e7454","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":134.0,"rawValue":1417.2421356877353,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a175a8ab65c8640f69817e77f12994d0ffa943f4","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":892,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=891","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":892,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1417.2421356877353,"rating_lower":1405.4816339095553,"rating_upper":1429.002637465915,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.00439534311673,"vote_count":2455.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1405.4816339095553,"type":"confidence_interval","upper":1429.002637465915}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1417.2421356877353,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8159af5a2d92f2afb3fe","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9ae1a6a71da6bac1fd313e78","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":144.0,"rawValue":1394.6498126524996,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a175a8ab65c8640f69817e77f12994d0ffa943f4","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":144,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=143","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":144,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1394.6498126524996,"rating_lower":1388.4582235485852,"rating_upper":1400.8414017564141,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.97948368604006,"vote_count":10982.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1388.4582235485852,"type":"confidence_interval","upper":1400.8414017564141}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1394.6498126524996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a16afa8b34e12a02f878","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-149ab517a4ede50b669b838d","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":64.0,"rawValue":1464.5854192460758,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d042c06cbd3352ee821d541b47443b84ccb3150","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":822,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=821","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":822,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1464.5854192460758,"rating_lower":1455.8024530659713,"rating_upper":1473.36838542618,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.08104174000166,"vote_count":5537.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1455.8024530659713,"type":"confidence_interval","upper":1473.36838542618}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.5854192460758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ce0935ec5be671b252fd","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-36271f70b23971bc94ff243b","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":81.0,"rawValue":1467.756636526635,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d042c06cbd3352ee821d541b47443b84ccb3150","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":475,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=474","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":475,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1467.756636526635,"rating_lower":1447.1610098349613,"rating_upper":1488.3522632183087,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":110.42155041144856,"vote_count":862.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.1610098349613,"type":"confidence_interval","upper":1488.3522632183087}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1467.756636526635,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d3d63ff9a50fc61f69bc","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f91b094976c08fd0671f8894","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":94.0,"rawValue":1422.2637647204183,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d042c06cbd3352ee821d541b47443b84ccb3150","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":94,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=93","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":94,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1422.2637647204183,"rating_lower":1416.575995894722,"rating_upper":1427.9515335461147,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.421465834877313,"vote_count":19434.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1416.575995894722,"type":"confidence_interval","upper":1427.9515335461147}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1422.2637647204183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b7e84e8945043798d9a7","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-78c9aed4a1c111d3e8c6fc19","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45.0,"rawValue":1476.424896591302,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30358d21032773c73451f5a26d080bc71171eacc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":803,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=802","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":803,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1476.424896591302,"rating_lower":1468.7070762401615,"rating_upper":1484.1427169424426,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.505763240673202,"vote_count":6913.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1468.7070762401615,"type":"confidence_interval","upper":1484.1427169424426}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1476.424896591302,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-338ac06ee750d1a0f408","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7beedab01aca0bb16e3a0fd","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72.0,"rawValue":1436.526656194733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30358d21032773c73451f5a26d080bc71171eacc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":72,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=71","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":72,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1436.526656194733,"rating_lower":1431.787799218506,"rating_upper":1441.2655131709603,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.845895137586903,"vote_count":24397.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1431.787799218506,"type":"confidence_interval","upper":1441.2655131709603}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1436.526656194733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e4716601440d03edbd5e","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-be5cfdd2174e7352ea6e8f83","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":105.0,"rawValue":1455.5304295602677,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30358d21032773c73451f5a26d080bc71171eacc","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":499,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=498","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":499,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1455.5304295602677,"rating_lower":1438.1259234574848,"rating_upper":1472.9349356630507,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":78.85463487203931,"vote_count":1284.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1438.1259234574848,"type":"confidence_interval","upper":1472.9349356630507}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1455.5304295602677,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-118e6a015a9ecf952509","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d0e99b2c36455efe5857c514","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":56.0,"rawValue":1468.9854695540266,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ccbe2eaaeda82574e450a580fb237907993d9cb0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":814,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=813","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":814,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1468.9854695540266,"rating_lower":1462.350395318017,"rating_upper":1475.6205437900364,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.460284275390032,"vote_count":12875.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1462.350395318017,"type":"confidence_interval","upper":1475.6205437900364}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1468.9854695540266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6032c8b48ecfa3ffd590","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3cf9a59e905a3d7455164c6e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":73.0,"rawValue":1471.3315083707973,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ccbe2eaaeda82574e450a580fb237907993d9cb0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":467,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=466","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":467,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1471.3315083707973,"rating_lower":1457.7885798425646,"rating_upper":1484.87443689903,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":47.745120195685566,"vote_count":2144.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.7885798425646,"type":"confidence_interval","upper":1484.87443689903}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1471.3315083707973,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6278dbbf68af6a2cef71","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-afa14db6221d376b9a23170d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":86.0,"rawValue":1427.3345994944987,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ccbe2eaaeda82574e450a580fb237907993d9cb0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":86,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=85","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":86,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1427.3345994944987,"rating_lower":1422.99736813579,"rating_upper":1431.671830853207,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.896987508397753,"vote_count":44557.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1422.99736813579,"type":"confidence_interval","upper":1431.671830853207}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1427.3345994944987,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f343dfb44ef3a538b8f7","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-24989972b2c048ddb00da5f6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":13.0,"rawValue":1499.3478828199889,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4224419b26329420dce70753be80111b189bf9ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":771,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=770","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":771,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1499.3478828199889,"rating_lower":1493.1299742535684,"rating_upper":1505.5657913864093,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.064506414096957,"vote_count":14938.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1493.1299742535684,"type":"confidence_interval","upper":1505.5657913864093}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1499.3478828199889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ff889bd9a1fe5c1bd2e8","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0421a9f2a337516c6f26f816","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28.0,"rawValue":1464.96741189912,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4224419b26329420dce70753be80111b189bf9ed","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":28,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=27","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":28,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1464.96741189912,"rating_lower":1460.9471567980147,"rating_upper":1468.9876670002252,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.207373248645847,"vote_count":54177.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.9471567980147,"type":"confidence_interval","upper":1468.9876670002252}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.96741189912,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a0ee6ec724946d72c512","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8bc84b9c4388ebbf7e20eb30","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":38.0,"rawValue":1504.1336126761466,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4224419b26329420dce70753be80111b189bf9ed","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":432,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=431","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":432,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1504.1336126761466,"rating_lower":1492.2712349033047,"rating_upper":1515.9959904489886,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.63087722496708,"vote_count":2924.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1492.2712349033047,"type":"confidence_interval","upper":1515.9959904489886}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1504.1336126761466,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c8a41a1e74c80a03b392","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-685283696547c9d6612288b8","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":128.0,"rawValue":1423.0252338599134,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9ae9dafddad859fb1944460431ddf22e1c6d4d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":886,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=885","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":886,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1423.0252338599134,"rating_lower":1413.0661141619826,"rating_upper":1432.984353557844,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.81937482276327,"vote_count":3436.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1413.0661141619826,"type":"confidence_interval","upper":1432.984353557844}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1423.0252338599134,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3b651177d127ee3fbe58","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c019694e961d2abb3a01a786","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":132.0,"rawValue":1430.9727612173845,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9ae9dafddad859fb1944460431ddf22e1c6d4d7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":526,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=525","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":526,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1430.9727612173845,"rating_lower":1411.1349663089911,"rating_upper":1450.810556125778,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":102.44496301964114,"vote_count":852.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1411.1349663089911,"type":"confidence_interval","upper":1450.810556125778}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1430.9727612173845,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf84b76cd3e62ea9bf8b","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e506d1975fbf75f9b6d3abce","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":148.0,"rawValue":1391.0609391665223,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9ae9dafddad859fb1944460431ddf22e1c6d4d7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":148,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=147","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":148,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1391.0609391665223,"rating_lower":1385.9524212476645,"rating_upper":1396.16945708538,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.793501257049904,"vote_count":17151.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1385.9524212476645,"type":"confidence_interval","upper":1396.16945708538}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1391.0609391665223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9e47e5e4344e6b9c8a65","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a210c98d0a34a6491809618f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":163.0,"rawValue":1388.6823361591046,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7575b5d211281eeff235b3f1eeb66d542ef133e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":557,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=556","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":557,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1388.6823361591046,"rating_lower":1375.2264623153346,"rating_upper":1402.1382100028748,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":47.13327653652593,"vote_count":2183.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1375.2264623153346,"type":"confidence_interval","upper":1402.1382100028748}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1388.6823361591046,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-63054f09d580900925e1","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b65c2a0f09f5930da48f04aa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":176.0,"rawValue":1378.5858443722477,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7575b5d211281eeff235b3f1eeb66d542ef133e2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":934,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=933","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":934,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1378.5858443722477,"rating_lower":1372.0215841884817,"rating_upper":1385.1501045560137,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.216965681904703,"vote_count":10842.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1372.0215841884817,"type":"confidence_interval","upper":1385.1501045560137}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1378.5858443722477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-70bd8ddcecede2857a78","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-32e37231e713d0cf59731729","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":178.0,"rawValue":1358.9609122667448,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7575b5d211281eeff235b3f1eeb66d542ef133e2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":178,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=177","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":178,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1358.9609122667448,"rating_lower":1354.9998107417878,"rating_upper":1362.9220137917018,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.084470515860321,"vote_count":41032.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1354.9998107417878,"type":"confidence_interval","upper":1362.9220137917018}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1358.9609122667448,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fe7307462f238941f150","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-893fc63bd9f9abe0274510af","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":78.0,"rawValue":1453.4669538933076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9f2227e97c738367e96f00aa3cc0e70b7929a64","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":836,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=835","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":836,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1453.4669538933076,"rating_lower":1447.7894286310875,"rating_upper":1459.144479155528,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.39115935058336,"vote_count":17908.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.7894286310875,"type":"confidence_interval","upper":1459.144479155528}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1453.4669538933076,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dadafe983390d0127123","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c99e6370ba6fa74ac23a774e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":125.0,"rawValue":1437.4985997127415,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9f2227e97c738367e96f00aa3cc0e70b7929a64","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":519,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=518","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":519,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1437.4985997127415,"rating_lower":1426.3557979994887,"rating_upper":1448.6414014259942,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.32158297571792,"vote_count":3376.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1426.3557979994887,"type":"confidence_interval","upper":1448.6414014259942}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1437.4985997127415,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-31006129e4f0e361f903","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6928a65c850186392c18b82f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":132.0,"rawValue":1405.2729700708335,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9f2227e97c738367e96f00aa3cc0e70b7929a64","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":132,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=131","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":132,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1405.2729700708335,"rating_lower":1401.5203714145534,"rating_upper":1409.0255687271133,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.665793994524611,"vote_count":62076.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1401.5203714145534,"type":"confidence_interval","upper":1409.0255687271133}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1405.2729700708335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-627c6518968bdd9b1f54","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ff5b4bd466261f0dcfea9d1b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":49.0,"rawValue":1473.4411759865222,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7be253a86a8ae3c265a1cb6e71255f91bfaf128","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":807,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=806","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":807,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1473.4411759865222,"rating_lower":1466.7816449350594,"rating_upper":1480.100707037985,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.544924960925604,"vote_count":12012.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1466.7816449350594,"type":"confidence_interval","upper":1480.100707037985}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1473.4411759865222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ee557acd0d49ce20023d","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-de7fae8ae533cfef859dc3ee","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":76.0,"rawValue":1434.7629318817199,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7be253a86a8ae3c265a1cb6e71255f91bfaf128","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":76,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=75","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":76,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1434.7629318817199,"rating_lower":1430.2820353925927,"rating_upper":1439.243828370847,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.226773026463067,"vote_count":41023.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.2820353925927,"type":"confidence_interval","upper":1439.243828370847}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1434.7629318817199,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e4e2aea2dc21b561bf30","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c651a233f07e3f9586420a18","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":98.0,"rawValue":1458.0451634068265,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7be253a86a8ae3c265a1cb6e71255f91bfaf128","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":492,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=491","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":492,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1458.0451634068265,"rating_lower":1444.3861750143324,"rating_upper":1471.7041517993207,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.56695662107371,"vote_count":2195.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1444.3861750143324,"type":"confidence_interval","upper":1471.7041517993207}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1458.0451634068265,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d229326733b53aa2aedd","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7ed15d96a7ac1213e3f73d61","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":142.0,"rawValue":1412.5711018740342,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df8c977c5df5a9c46a067c927e4cbf6e19f020d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":900,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=899","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":900,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1412.5711018740342,"rating_lower":1403.637330861949,"rating_upper":1421.5048728861195,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.776550842201342,"vote_count":4869.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1403.637330861949,"type":"confidence_interval","upper":1421.5048728861195}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1412.5711018740342,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5ef465c8e5c8e1bde3d1","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-95e4f11abe79909efc5f2cbf","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":181.0,"rawValue":1356.5686108842956,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df8c977c5df5a9c46a067c927e4cbf6e19f020d4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":181,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=180","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":181,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1356.5686108842956,"rating_lower":1351.610443815558,"rating_upper":1361.5267779530332,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.39950129078084,"vote_count":25788.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1351.610443815558,"type":"confidence_interval","upper":1361.5267779530332}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1356.5686108842956,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4a7a73ca14539c84d7b1","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2abf0829b8f18f0d99bd3cf4","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":187.0,"rawValue":1355.1825086820477,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df8c977c5df5a9c46a067c927e4cbf6e19f020d4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":581,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=580","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":581,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1355.1825086820477,"rating_lower":1338.4759758184114,"rating_upper":1371.889041545684,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":72.65683516381547,"vote_count":1352.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1338.4759758184114,"type":"confidence_interval","upper":1371.889041545684}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1355.1825086820477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bdf77257f40fa1f1f7da","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6ea801bc49e01e96ed78f9fb","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":91.0,"rawValue":1463.164696604873,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3cbdde4fcd1576180acb8130f5da173518cc3b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":485,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=484","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":485,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1463.164696604873,"rating_lower":1448.1007651322411,"rating_upper":1478.2286280775045,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":59.0718375502825,"vote_count":1649.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.1007651322411,"type":"confidence_interval","upper":1478.2286280775045}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1463.164696604873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7d4d5ce5812a340cf1d7","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-76420184ec102a973ea9ecb0","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":108.0,"rawValue":1436.2856315236722,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3cbdde4fcd1576180acb8130f5da173518cc3b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":866,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=865","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":866,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1436.2856315236722,"rating_lower":1429.056749835994,"rating_upper":1443.5145132113503,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.603355624415288,"vote_count":7796.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1429.056749835994,"type":"confidence_interval","upper":1443.5145132113503}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1436.2856315236722,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6f8de918f5aba67833ff","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6067f2b718e5161a4354b4a0","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":108.0,"rawValue":1417.8843322849618,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3cbdde4fcd1576180acb8130f5da173518cc3b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":108,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=107","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":108,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1417.8843322849618,"rating_lower":1413.535498256519,"rating_upper":1422.2331663134044,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.923222736388152,"vote_count":28413.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1413.535498256519,"type":"confidence_interval","upper":1422.2331663134044}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1417.8843322849618,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-67e2f661938031b2e75e","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e000cd7f9c8595d2b2ea44a4","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":66.0,"rawValue":1475.397041496677,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a8bc1bda534f6e9f43bf8179e5e5d7f72a496e4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":460,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=459","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":460,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1475.397041496677,"rating_lower":1459.7740107629327,"rating_upper":1491.0200722304214,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":63.53812462928328,"vote_count":1525.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1459.7740107629327,"type":"confidence_interval","upper":1491.0200722304214}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1475.397041496677,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7ad27c2f032ad167677d","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e5eab78ada5f861875120ee9","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":125.0,"rawValue":1424.5979855989967,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a8bc1bda534f6e9f43bf8179e5e5d7f72a496e4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":883,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=882","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":883,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1424.5979855989967,"rating_lower":1417.3107623853741,"rating_upper":1431.8852088126193,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.82381658735695,"vote_count":7511.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1417.3107623853741,"type":"confidence_interval","upper":1431.8852088126193}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1424.5979855989967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-98bc3290f349f8460781","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d001f427c427250d4fee3510","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":128.0,"rawValue":1407.8616630235158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a8bc1bda534f6e9f43bf8179e5e5d7f72a496e4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":128,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=127","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":128,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1407.8616630235158,"rating_lower":1403.4787919620371,"rating_upper":1412.2445340849945,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.000589525537647,"vote_count":27282.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1403.4787919620371,"type":"confidence_interval","upper":1412.2445340849945}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1407.8616630235158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9e165cd5f6026936bdb7","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-85bc6854e9d0d779970fb6d5","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":90.0,"rawValue":1464.5961634775433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b22ba5dca6574c234cc1b987ac6a8f44123bff2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":484,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=483","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":484,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1464.5961634775433,"rating_lower":1449.399342507048,"rating_upper":1479.7929844480389,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":60.11866282808638,"vote_count":1620.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1449.399342507048,"type":"confidence_interval","upper":1479.7929844480389}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.5961634775433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-841990f35f0d2c79c36d","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-77569fbb7fa8c07d851250b0","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":142.0,"rawValue":1395.586964212609,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b22ba5dca6574c234cc1b987ac6a8f44123bff2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":142,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=141","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":142,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1395.586964212609,"rating_lower":1391.2889433804926,"rating_upper":1399.8849850447257,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.808845788946945,"vote_count":29097.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1391.2889433804926,"type":"confidence_interval","upper":1399.8849850447257}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1395.586964212609,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d4c3162bd4df1fe40468","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f9bb31ab9622a6e3729286a0","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":144.0,"rawValue":1409.761729852643,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b22ba5dca6574c234cc1b987ac6a8f44123bff2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":902,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=901","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":902,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1409.761729852643,"rating_lower":1402.5674910792884,"rating_upper":1416.9559686259975,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.473285526118337,"vote_count":7973.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1402.5674910792884,"type":"confidence_interval","upper":1416.9559686259975}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1409.761729852643,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e4804538047e81741637","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8a7041aa9a6e4f8a87dc58a4","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45.0,"rawValue":1498.948733077815,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e9cabca852262bdadf7181ba21d0b04897f250c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":439,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=438","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":439,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1498.948733077815,"rating_lower":1488.4985158133047,"rating_upper":1509.3989503423252,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.42853352668128,"vote_count":3861.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1488.4985158133047,"type":"confidence_interval","upper":1509.3989503423252}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1498.948733077815,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-910c03695cca11fc07d1","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-21cd79b4c6d36d087531b86f","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63.0,"rawValue":1464.7670373393412,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e9cabca852262bdadf7181ba21d0b04897f250c8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":821,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=820","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":821,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1464.7670373393412,"rating_lower":1459.4519079567713,"rating_upper":1470.082166721911,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.354133331137012,"vote_count":20247.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1459.4519079567713,"type":"confidence_interval","upper":1470.082166721911}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1464.7670373393412,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-66795d81222f3b581dbc","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1acd3f339c4b4a09876a2000","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":67.0,"rawValue":1438.2532016582525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e9cabca852262bdadf7181ba21d0b04897f250c8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":67,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=66","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":67,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1438.2532016582525,"rating_lower":1434.8315444085888,"rating_upper":1441.6748589079164,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.0477323539449546,"vote_count":70207.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1434.8315444085888,"type":"confidence_interval","upper":1441.6748589079164}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1438.2532016582525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ed015edd05189e121b76","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f6ce8cb49b4f758035419399","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":111.0,"rawValue":1451.0570878730985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"96eb968733577ef3fe917709e642eee359988aab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":505,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=504","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":505,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1451.0570878730985,"rating_lower":1439.4864563327078,"rating_upper":1462.6277194134889,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.851216814368854,"vote_count":2937.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1439.4864563327078,"type":"confidence_interval","upper":1462.6277194134889}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1451.0570878730985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a2e97fc9c7053c31b09f","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8957ec3e6a40e8ef6e87d9b0","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":140.0,"rawValue":1397.599796544761,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"96eb968733577ef3fe917709e642eee359988aab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":140,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=139","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":140,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1397.599796544761,"rating_lower":1393.8480084350442,"rating_upper":1401.3515846544776,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":3.6642105713547775,"vote_count":58319.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1393.8480084350442,"type":"confidence_interval","upper":1401.3515846544776}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1397.599796544761,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ab29387f200ed3239c9e","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7a596923be54022ee8744fc","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":141.0,"rawValue":1412.8720926205697,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"96eb968733577ef3fe917709e642eee359988aab","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":899,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=898","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":899,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1412.8720926205697,"rating_lower":1407.083985937858,"rating_upper":1418.6601993032814,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.721212574237224,"vote_count":16630.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1407.083985937858,"type":"confidence_interval","upper":1418.6601993032814}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1412.8720926205697,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-41cd56bca0594bf808c3","canonicalModelId":"qwen/qwen3.5-max-preview@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-feb71eb534cf165a4e5be1da","mappingCandidates":["qwen/qwen3.5-max-preview@2026-03-18"],"mappingEvidence":{"confidence":"high","note":"Arena explicit Qwen3.5 Max Preview endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena explicit Qwen3.5 Max Preview endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Max Preview","modelRef":"qwen3.5-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11.0,"rawValue":1535.6774423499453,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5df640f6fc65f05cc9d07dd3f4adabf9dd8452b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":405,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=404","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":405,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1535.6774423499453,"rating_lower":1518.2172478946213,"rating_upper":1553.137636805269,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":79.3600568553375,"vote_count":1269.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1518.2172478946213,"type":"confidence_interval","upper":1553.137636805269}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1535.6774423499453,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-084818de2d5771751392","canonicalModelId":"qwen/qwen3.5-max-preview@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3449886053bb184c57112a88","mappingCandidates":["qwen/qwen3.5-max-preview@2026-03-18"],"mappingEvidence":{"confidence":"high","note":"Arena explicit Qwen3.5 Max Preview endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena explicit Qwen3.5 Max Preview endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Max Preview","modelRef":"qwen3.5-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21.0,"rawValue":1470.9246252658847,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5df640f6fc65f05cc9d07dd3f4adabf9dd8452b3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":21,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=20","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":21,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1470.9246252658847,"rating_lower":1465.9765423314418,"rating_upper":1475.8727082003277,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.37349659827979,"vote_count":21484.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1465.9765423314418,"type":"confidence_interval","upper":1475.8727082003277}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1470.9246252658847,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ba35a14c4dcf84db0d28","canonicalModelId":"qwen/qwen3.5-max-preview@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-86f8b9e5fe9631be2305c4ee","mappingCandidates":["qwen/qwen3.5-max-preview@2026-03-18"],"mappingEvidence":{"confidence":"high","note":"Arena explicit Qwen3.5 Max Preview endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena explicit Qwen3.5 Max Preview endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Max Preview","modelRef":"qwen3.5-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29.0,"rawValue":1486.9869932732572,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5df640f6fc65f05cc9d07dd3f4adabf9dd8452b3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":787,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=786","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":787,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1486.9869932732572,"rating_lower":1478.9378288055245,"rating_upper":1495.0361577409901,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.865740764834275,"vote_count":6055.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1478.9378288055245,"type":"confidence_interval","upper":1495.0361577409901}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1486.9869932732572,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6dd02e2e4b0a927184d0","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-06c8275cb7f0e88f385125a7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":46.0,"rawValue":1446.2675776614128,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de12e0168afbc7ee9ea67fb9f506d1179f03e9da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":46,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=45","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":46,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1446.2675776614128,"rating_lower":1437.9346082590055,"rating_upper":1454.6005470638202,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.076044102669382,"vote_count":5196.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1437.9346082590055,"type":"confidence_interval","upper":1454.6005470638202}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1446.2675776614128,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e9433b152604d9980aa9","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6701af42ce2324130c84fe56","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":52.0,"rawValue":1490.283888090373,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de12e0168afbc7ee9ea67fb9f506d1179f03e9da","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":446,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=445","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":446,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1490.283888090373,"rating_lower":1457.531183416849,"rating_upper":1523.036592763897,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":279.2531987202839,"vote_count":319.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.531183416849,"type":"confidence_interval","upper":1523.036592763897}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1490.283888090373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ca317b9a24541d00c589","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-98c06a46043d15221a6dab5a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":59.0,"rawValue":1467.9806594898603,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de12e0168afbc7ee9ea67fb9f506d1179f03e9da","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":817,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=816","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":817,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1467.9806594898603,"rating_lower":1452.5689651812322,"rating_upper":1483.3923537984883,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":61.830760799273634,"vote_count":1551.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1452.5689651812322,"type":"confidence_interval","upper":1483.3923537984883}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1467.9806594898603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9db7320ac52433fdedee","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-08f7b68788879ff60a204e68","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":62.0,"rawValue":1465.9333014962801,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"104d90349294f6c7ade64b6bacb0a08dfb0eef18","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":820,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=819","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":820,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1465.9333014962801,"rating_lower":1459.5651833827944,"rating_upper":1472.301419609766,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.556647929907607,"vote_count":13781.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1459.5651833827944,"type":"confidence_interval","upper":1472.301419609766}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1465.9333014962801,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4c2e542512955810436c","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-171f7eefb88e0f5f028425cb","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":68.0,"rawValue":1474.8119844446103,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"104d90349294f6c7ade64b6bacb0a08dfb0eef18","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":462,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=461","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":462,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1474.8119844446103,"rating_lower":1461.5640249114708,"rating_upper":1488.05994397775,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":45.68796386576665,"vote_count":2224.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1461.5640249114708,"type":"confidence_interval","upper":1488.05994397775}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1474.8119844446103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-64d961344d9c63d20992","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c7ba10f12b9a26b992deec6f","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":70.0,"rawValue":1436.8284922613113,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"104d90349294f6c7ade64b6bacb0a08dfb0eef18","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":70,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=69","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":70,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1436.8284922613113,"rating_lower":1432.6073693884928,"rating_upper":1441.0496151341301,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":4.6383103761121145,"vote_count":45376.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1432.6073693884928,"type":"confidence_interval","upper":1441.0496151341301}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1436.8284922613113,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3203c6ebb3f71b5991a7","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-091145f24182ed7e5a9e46bd","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":16.0,"rawValue":1497.5206352718767,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82cf534266119702553eda337ed202df475ffffc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":774,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=773","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":774,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1497.5206352718767,"rating_lower":1479.1653257921655,"rating_upper":1515.8759447515877,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":87.70558317089925,"vote_count":1119.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1479.1653257921655,"type":"confidence_interval","upper":1515.8759447515877}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1497.5206352718767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1bf16ee17b63a60cd24a","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f3bdc729fcb255546c357c29","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":19.0,"rawValue":1474.2960161967428,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82cf534266119702553eda337ed202df475ffffc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":19,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=18","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":19,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1474.2960161967428,"rating_lower":1464.3540662917903,"rating_upper":1484.2379661016955,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.730424957341395,"vote_count":3712.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1464.3540662917903,"type":"confidence_interval","upper":1484.2379661016955}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1474.2960161967428,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e857361e214dcc762ad5","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-deb42559be97f11c66c12929","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21.0,"rawValue":1525.6221592269271,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82cf534266119702553eda337ed202df475ffffc","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":415,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=414","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":415,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1525.6221592269271,"rating_lower":1488.562833016711,"rating_upper":1562.681485437143,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":357.5187769179423,"vote_count":264.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1488.562833016711,"type":"confidence_interval","upper":1562.681485437143}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1525.6221592269271,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2268b2a59250f25d0975","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-03f8b2e206df5847da0f8840","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":33.0,"rawValue":1456.2401150516928,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e143397b2b684823996a2fae57ed3ce6a5fa831e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":33,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=32","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":33,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1456.2401150516928,"rating_lower":1451.5782170308407,"rating_upper":1460.9020130725448,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.657562444700935,"vote_count":34426.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1451.5782170308407,"type":"confidence_interval","upper":1460.9020130725448}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1456.2401150516928,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-78340d005a333a4eccc1","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cf3b8e98f05a5c08f6738ade","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":34.0,"rawValue":1508.2299700220494,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e143397b2b684823996a2fae57ed3ce6a5fa831e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":428,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=427","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":428,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1508.2299700220494,"rating_lower":1493.9384528648086,"rating_upper":1522.5214871792903,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":53.16924433900437,"vote_count":1901.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1493.9384528648086,"type":"confidence_interval","upper":1522.5214871792903}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1508.2299700220494,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a12f9933fa1e3f596c43","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ce242a39a3d06be5babaae48","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":47.0,"rawValue":1475.3522024096037,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e143397b2b684823996a2fae57ed3ce6a5fa831e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":805,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=804","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":805,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1475.3522024096037,"rating_lower":1468.464597931382,"rating_upper":1482.2398068878254,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.349239615133358,"vote_count":10112.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1468.464597931382,"type":"confidence_interval","upper":1482.2398068878254}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1475.3522024096037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fa1107ee8505638bb902","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-43ed5188543944ce29893cc9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":39.0,"rawValue":1482.767837213196,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7441645d44dccc927959c6787cf5bd438a292521","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":797,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=796","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":797,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1482.767837213196,"rating_lower":1463.6041332300117,"rating_upper":1501.9315411963798,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":95.60106394391798,"vote_count":950.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1463.6041332300117,"type":"confidence_interval","upper":1501.9315411963798}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1482.767837213196,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c15a179cf94f8eb5b6ec","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c0f62f3c1c154b11ce4d9ddb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":61.0,"rawValue":1440.7741507579944,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7441645d44dccc927959c6787cf5bd438a292521","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":61,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=60","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":61,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1440.7741507579944,"rating_lower":1430.3159074617424,"rating_upper":1451.2323940542462,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.472217964277302,"vote_count":3205.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.3159074617424,"type":"confidence_interval","upper":1451.2323940542462}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1440.7741507579944,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5daf31e1bc51fbfd6d54","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a13a6d7b45dc0cbb722be80a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6.0,"rawValue":1545.239658145461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adc70e3b78af346ddebcb3547d67c25220115c7b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":400,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=399","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":400,"metadata":{"category":"chinese","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1545.239658145461,"rating_lower":1521.247541375223,"rating_upper":1569.2317749156987,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":149.8445496840489,"vote_count":621.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1521.247541375223,"type":"confidence_interval","upper":1569.2317749156987}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1545.239658145461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-13bac39d2638af3d1bff","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-be4641c77c983cc0fea1e314","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":11.0,"rawValue":1481.889534951889,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adc70e3b78af346ddebcb3547d67c25220115c7b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":11,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=10","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":11,"metadata":{"category":"overall","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1481.889534951889,"rating_lower":1475.1022440594659,"rating_upper":1488.6768258443117,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.992141477659526,"vote_count":9955.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1475.1022440594659,"type":"confidence_interval","upper":1488.6768258443117}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1481.889534951889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-text","benchmarkName":"Arena Text","benchmarkVersion":null,"benchmarkVersionHint":"arena-text@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4ff2dba7e675f16057c1","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3b8e4361498015d8ed0d14cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"text","category":"coding","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":14.0,"rawValue":1499.1631125049637,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adc70e3b78af346ddebcb3547d67c25220115c7b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":772,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=text;split=latest;row_idx=771","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":772,"metadata":{"category":"coding","config":"text","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1499.1631125049637,"rating_lower":1488.1448160472266,"rating_upper":1510.1814089627007,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.603321159288914,"vote_count":2970.0},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1488.1448160472266,"type":"confidence_interval","upper":1510.1814089627007}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1499.1631125049637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-aca36c9637975645804e","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fffcaaa4edf3494b9bf5f642","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1366.0454732568385,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1410,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=409","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1410,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1366.0454732568385,"rating_lower":1353.662909817285,"rating_upper":1378.428036696392,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.91397135603251,"vote_count":2647},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1353.662909817285,"type":"confidence_interval","upper":1378.428036696392}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1366.0454732568385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2790c97493004a27c44e","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c52ac83e98f4be3a487e477f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1328.0663081312318,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1001,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=0","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1001,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1328.0663081312318,"rating_lower":1319.4735786431747,"rating_upper":1336.659037619289,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.22056268237799,"vote_count":9562},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1319.4735786431747,"type":"confidence_interval","upper":1336.659037619289}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1328.0663081312318,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7b76138f3069697f4ff7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bd2cb0398b9ad05ee08ab84b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1356.2812075554702,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1707,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=706","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1707,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1356.2812075554702,"rating_lower":1337.7684550116855,"rating_upper":1374.793960099255,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":89.2166290840201,"vote_count":1124},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1337.7684550116855,"type":"confidence_interval","upper":1374.793960099255}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1356.2812075554702,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-faff3368d5042e3a8aad","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-23060da405f4f69623b3d036","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1329.250814778743,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1513,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=512","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1513,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1329.250814778743,"rating_lower":1317.991704447266,"rating_upper":1340.5099251102201,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.99985015418338,"vote_count":4059},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1317.991704447266,"type":"confidence_interval","upper":1340.5099251102201}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1329.250814778743,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f14897ed998ed61277ba","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-02bb5d29d8a00bab3f09afce","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1339.4591306887673,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1885,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=884","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1885,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1339.4591306887673,"rating_lower":1330.6294993657193,"rating_upper":1348.2887620118156,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.294995453541457,"vote_count":6864},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1330.6294993657193,"type":"confidence_interval","upper":1348.2887620118156}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1339.4591306887673,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-22ff58f4ef3bb8d197d4","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2277d8e9bd5e881a93d68473","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1339.609797891108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1325,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=324","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1325,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1339.609797891108,"rating_lower":1315.1101145188845,"rating_upper":1364.1094812633316,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":156.2517036771862,"vote_count":632},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1315.1101145188845,"type":"confidence_interval","upper":1364.1094812633316}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1339.609797891108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1312a5244bf3b0def92d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3724a34e8fe3910ac14115a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1330.367604811887,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1807,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=806","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1807,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1330.367604811887,"rating_lower":1297.0751856638674,"rating_upper":1363.6600239599068,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":288.5323582688192,"vote_count":316},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1297.0751856638674,"type":"confidence_interval","upper":1363.6600239599068}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1330.367604811887,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e0884349988aa0653207","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ef7ef9de0eff99fb5c1ce602","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1386.2439096452076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef2c8da2bda142476a1e7fe66d505aca8be139","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1183,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=182","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1183,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1386.2439096452076,"rating_lower":1356.3614082035879,"rating_upper":1416.1264110868274,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":232.4543706151337,"vote_count":459},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1356.3614082035879,"type":"confidence_interval","upper":1416.1264110868274}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1386.2439096452076,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-32f2311e535baa1a9a30","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8522d0d383fbaaa00a5689f5","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1253.4276980479322,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1298,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=297","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1298,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1253.4276980479322,"rating_lower":1220.7701217028475,"rating_upper":1286.0852743930168,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":277.63340504645186,"vote_count":192},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1220.7701217028475,"type":"confidence_interval","upper":1286.0852743930168}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1253.4276980479322,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-88593eac7ae4fab1d1a4","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-55a9268e34a69a244341db36","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72,"rawValue":1224.2350785852213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1778,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=777","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1778,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1224.2350785852213,"rating_lower":1192.6071660959724,"rating_upper":1255.8629910744698,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":260.402335445789,"vote_count":327},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1192.6071660959724,"type":"confidence_interval","upper":1255.8629910744698}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1224.2350785852213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ba3c5044d13de9dbdd01","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aecdc1d5125754b8a43209fd","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":84,"rawValue":1178.450043786878,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1968,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=967","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1968,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1178.450043786878,"rating_lower":1162.2744373548853,"rating_upper":1194.625650218871,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":68.11220831867445,"vote_count":1351},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1162.2744373548853,"type":"confidence_interval","upper":1194.625650218871}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1178.450043786878,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b5c2674d3821e7f78363","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b2674af7af509454f9599368","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":86,"rawValue":1175.5783200444741,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1495,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=494","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1495,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1175.5783200444741,"rating_lower":1150.147132769157,"rating_upper":1201.0095073197913,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":168.35929172225138,"vote_count":465},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1150.147132769157,"type":"confidence_interval","upper":1201.0095073197913}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1175.5783200444741,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ab1256b567ff77f028a5","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8a5840fbc64250c20cd2b1ab","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":86,"rawValue":1186.696775338311,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1598,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=597","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1598,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1186.696775338311,"rating_lower":1168.7716418289642,"rating_upper":1204.6219088476578,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":83.64281027743762,"vote_count":1029},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1168.7716418289642,"type":"confidence_interval","upper":1204.6219088476578}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1186.696775338311,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3e81bb8359f795040b62","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-08c3ade848d7a4bcee3d4675","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":88,"rawValue":1175.4214503141313,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1088,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=87","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1088,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1175.4214503141313,"rating_lower":1162.5466228675814,"rating_upper":1188.2962777606815,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.15058146282177,"vote_count":2201},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1162.5466228675814,"type":"confidence_interval","upper":1188.2962777606815}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1175.4214503141313,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7f5134a556e245b2f6cc","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-515bbf66032a23291646b27d","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":77,"rawValue":1195.870868422259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1961,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=960","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1961,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1195.870868422259,"rating_lower":1175.8382076548935,"rating_upper":1215.9035291896246,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":104.46747346567444,"vote_count":843},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1175.8382076548935,"type":"confidence_interval","upper":1215.9035291896246}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1195.870868422259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fa9bfbdc54d9a299c497","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-362309543a4e5781d922676a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":79,"rawValue":1201.0411343597555,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1785,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=784","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1785,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1201.0411343597555,"rating_lower":1156.863381009719,"rating_upper":1245.218887709792,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":508.0553982624757,"vote_count":168},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1156.863381009719,"type":"confidence_interval","upper":1245.218887709792}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1201.0411343597555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8c751aed0b9f132bd408","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9d0dd3210e541cd607e74e90","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":80,"rawValue":1191.1950598829105,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1080,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=79","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1080,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1191.1950598829105,"rating_lower":1175.242103082324,"rating_upper":1207.1480166834972,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":66.25004784911188,"vote_count":1300},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1175.242103082324,"type":"confidence_interval","upper":1207.1480166834972}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1191.1950598829105,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a4706fd1083553830d4a","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fc7a0d7a6cfd8f5f1ab3b310","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":87,"rawValue":1175.315091204888,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1496,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=495","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1496,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1175.315091204888,"rating_lower":1140.7969018438635,"rating_upper":1209.8332805659122,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":310.17002976703407,"vote_count":282},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1140.7969018438635,"type":"confidence_interval","upper":1209.8332805659122}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1175.315091204888,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b79753b6b9a1d7077a9c","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0f30aec3cb97a88d33e0687a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Arena labels for the dated Claude Opus 4 release; thinking and context settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514-thinking-16k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":88,"rawValue":1183.6309405267036,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90210d115b0265f48d580ff1fc04075796676c10","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1600,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=599","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1600,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1183.6309405267036,"rating_lower":1159.8839440513595,"rating_upper":1207.377937002048,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":146.798356541568,"vote_count":590},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1159.8839440513595,"type":"confidence_interval","upper":1207.377937002048}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1183.6309405267036,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-64c4762f91f22c048185","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ac91dce894e08737886f2069","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1339.6258946871856,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1412,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=411","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1412,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1339.6258946871856,"rating_lower":1330.6694645970122,"rating_upper":1348.5823247773592,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.88207727960736,"vote_count":6827},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1330.6694645970122,"type":"confidence_interval","upper":1348.5823247773592}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1339.6258946871856,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ad841b6a9c9e4342816b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c050e3b10f609c664768d342","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1319.34049373008,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1809,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=808","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1809,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1319.34049373008,"rating_lower":1298.6301361935316,"rating_upper":1340.0508512666283,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":111.65521467549218,"vote_count":914},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1298.6301361935316,"type":"confidence_interval","upper":1340.0508512666283}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1319.34049373008,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-db5c525e492329b7d825","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-91f23ca3d2e46a109854aa64","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1325.267920995656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1889,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=888","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1889,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1325.267920995656,"rating_lower":1318.8444129586642,"rating_upper":1331.6914290326476,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.74109015018446,"vote_count":18261},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1318.8444129586642,"type":"confidence_interval","upper":1331.6914290326476}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1325.267920995656,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4b50a31cd4321e99d1a7","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-26981c01b893a195e4006fa4","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1380.6075943954047,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1184,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=183","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1184,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1380.6075943954047,"rating_lower":1360.3831273066023,"rating_upper":1400.832061484207,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":106.47753577953044,"vote_count":1380},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1360.3831273066023,"type":"confidence_interval","upper":1400.832061484207}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1380.6075943954047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3865eb034693f69ccef6","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e57a2313ac40c64c4d008488","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1315.8592803719687,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1518,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=517","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1518,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1315.8592803719687,"rating_lower":1307.229838271467,"rating_upper":1324.48872247247,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.385154036989405,"vote_count":10897},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1307.229838271467,"type":"confidence_interval","upper":1324.48872247247}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1315.8592803719687,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e276e986abccfb504318","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-20718926b0347f1acf726c73","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1316.7408184183773,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1330,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=329","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1330,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1316.7408184183773,"rating_lower":1300.4913216245511,"rating_upper":1332.9903152122033,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":68.7359043471025,"vote_count":1488},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1300.4913216245511,"type":"confidence_interval","upper":1332.9903152122033}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1316.7408184183773,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1ac46419d55f0832805f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f33d570ea7b711ae1aabac3b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1311.3488129222949,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1008,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=7","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1008,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1311.3488129222949,"rating_lower":1304.8036560296089,"rating_upper":1317.8939698149807,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.151773518721017,"vote_count":25169},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1304.8036560296089,"type":"confidence_interval","upper":1317.8939698149807}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1311.3488129222949,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f0a4f16e91d5bdefcf0b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-516f530a84f8def022ed1943","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1328.281760751524,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1719,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=718","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1719,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1328.281760751524,"rating_lower":1317.2470810458499,"rating_upper":1339.316440457198,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.69737380780087,"vote_count":3576},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1317.2470810458499,"type":"confidence_interval","upper":1339.316440457198}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1328.281760751524,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-59d051a027dbb984f9f6","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-479ebf6561042b93e3574158","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1254.8661093976898,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1676,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=675","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1676,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1254.8661093976898,"rating_lower":1212.7257055019504,"rating_upper":1297.006513293429,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":462.2758497187665,"vote_count":161},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1212.7257055019504,"type":"confidence_interval","upper":1297.006513293429}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1254.8661093976898,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fe607b3ecc3f7839ad73","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-24147305a0adfe747520b423","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1331.6823510071824,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1326,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=325","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1326,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1331.6823510071824,"rating_lower":1313.6074460247694,"rating_upper":1349.7572559895955,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":85.04638611854884,"vote_count":1201},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1313.6074460247694,"type":"confidence_interval","upper":1349.7572559895955}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1331.6823510071824,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-489a540193c409a90cf1","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fb2f05f2669069bae4ab6810","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1325.798127104686,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1888,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=887","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1888,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1325.798127104686,"rating_lower":1319.0861537076917,"rating_upper":1332.5101005016804,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.727468387600966,"vote_count":14932},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1319.0861537076917,"type":"confidence_interval","upper":1332.5101005016804}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1325.798127104686,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b87eeb0eaee32ca0cd4e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-212acbdbdbea960b7dd24c2d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1319.3019854373526,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1516,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=515","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1516,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1319.3019854373526,"rating_lower":1310.3093873626976,"rating_upper":1328.2945835120076,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.051070415398982,"vote_count":8897},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1310.3093873626976,"type":"confidence_interval","upper":1328.2945835120076}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1319.3019854373526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-895f8323cd1cdc249353","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9861f7a02fb8660c0826efa6","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1314.5661177451846,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1006,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=5","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1006,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1314.5661177451846,"rating_lower":1307.8911475633531,"rating_upper":1321.2410879270158,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.598517388320142,"vote_count":20867},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1307.8911475633531,"type":"confidence_interval","upper":1321.2410879270158}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1314.5661177451846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4be50224a2ad48040e74","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c75ea9f46c1a43407bb0eb62","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1308.616440051339,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1811,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=810","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1811,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1308.616440051339,"rating_lower":1286.3538074883645,"rating_upper":1330.8790726143131,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":129.0199457466654,"vote_count":770},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1286.3538074883645,"type":"confidence_interval","upper":1330.8790726143131}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1308.616440051339,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bd7d362e5a5efa4f1df9","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-654413bc529fe7f897ee313e","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1331.2485350549919,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1416,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=415","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1416,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1331.2485350549919,"rating_lower":1321.8269640727847,"rating_upper":1340.670106037199,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.10736725709014,"vote_count":5600},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1321.8269640727847,"type":"confidence_interval","upper":1340.670106037199}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1331.2485350549919,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-529dd6398d3492f36cff","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4d515c8cc99febee54875edf","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1372.6502759169039,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1187,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=186","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1187,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1372.6502759169039,"rating_lower":1351.785926825029,"rating_upper":1393.5146250087787,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":113.32180906964882,"vote_count":1209},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1351.785926825029,"type":"confidence_interval","upper":1393.5146250087787}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1372.6502759169039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8023f4ee921675cfe8d8","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-14cc2df196584d7e9c5d6efe","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1335.4145440824989,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1715,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=714","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1715,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1335.4145440824989,"rating_lower":1323.2377390779827,"rating_upper":1347.591349087015,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.59850828525004,"vote_count":2724},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1323.2377390779827,"type":"confidence_interval","upper":1347.591349087015}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1335.4145440824989,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d938ba9e862d7d578e09","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-60837682f69296def843425b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1277.7696033143948,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b9fbf1b48a79ef64b08b7948fc9137407b5a9db","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1669,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=668","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1669,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1277.7696033143948,"rating_lower":1228.161427368324,"rating_upper":1327.3777792604658,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":640.6345181780837,"vote_count":104},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1228.161427368324,"type":"confidence_interval","upper":1327.3777792604658}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1277.7696033143948,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-70a4aa936b0324fb169f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-da8386e63dde5ce1db6a3772","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1321.4273346920506,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1514,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=513","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1514,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1321.4273346920506,"rating_lower":1312.374240466197,"rating_upper":1330.4804289179044,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.335258006846587,"vote_count":9228},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1312.374240466197,"type":"confidence_interval","upper":1330.4804289179044}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1321.4273346920506,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-be24ac3c5eb8eff15a32","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d507c885400b90ab6dd7a046","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1316.5756460386099,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1003,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=2","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1003,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1316.5756460386099,"rating_lower":1309.7198007487484,"rating_upper":1323.4314913284713,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.23561590334102,"vote_count":21445},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1309.7198007487484,"type":"confidence_interval","upper":1323.4314913284713}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1316.5756460386099,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2b22b37565625b2e033c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-838ba333508b52f1ef802d32","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1325.846767304178,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1887,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=886","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1887,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1325.846767304178,"rating_lower":1319.0017505503213,"rating_upper":1332.6917840580347,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.196995086390974,"vote_count":15535},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1319.0017505503213,"type":"confidence_interval","upper":1332.6917840580347}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1325.846767304178,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-127cdb10212dcceff54b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ebc0aabe2073286ae57ca0ef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1336.5227447457085,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1414,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=413","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1414,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1336.5227447457085,"rating_lower":1327.143521416702,"rating_upper":1345.901968074715,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.900110182486188,"vote_count":5930},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1327.143521416702,"type":"confidence_interval","upper":1345.901968074715}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1336.5227447457085,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0b4ce543c86a1db26812","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0d0d366aa7fcaa2022bffe35","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1336.323477600447,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1713,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=712","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1713,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1336.323477600447,"rating_lower":1324.3736227850752,"rating_upper":1348.2733324158185,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":37.17312530833182,"vote_count":2973},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1324.3736227850752,"type":"confidence_interval","upper":1348.2733324158185}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1336.323477600447,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-df09c70827609a9428b4","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e92792b0b56ddf43c3c9b7fe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1311.0181433353632,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1332,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=331","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1332,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1311.0181433353632,"rating_lower":1293.0977738629754,"rating_upper":1328.9385128077513,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":83.59835599353563,"vote_count":1201},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1293.0977738629754,"type":"confidence_interval","upper":1328.9385128077513}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1311.0181433353632,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fc0270aca02d94d513ac","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a790974467dbb738b0f37d1a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1368.6043696266045,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1188,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=187","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1188,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1368.6043696266045,"rating_lower":1348.2425403635266,"rating_upper":1388.9661988896823,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":107.9288130606109,"vote_count":1255},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1348.2425403635266,"type":"confidence_interval","upper":1388.9661988896823}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1368.6043696266045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7a5da1472db707bd7080","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6ccb68513e7d60f246bab263","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1303.2734404857263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1814,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=813","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1814,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1303.2734404857263,"rating_lower":1281.8675882656728,"rating_upper":1324.6792927057797,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":119.28033870839953,"vote_count":785},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1281.8675882656728,"type":"confidence_interval","upper":1324.6792927057797}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1303.2734404857263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c6de0136448253e9a88d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-163694879faef97c3353b6da","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1319.8858391023387,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1515,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=514","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1515,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1319.8858391023387,"rating_lower":1310.8526831184201,"rating_upper":1328.9189950862572,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.241385327425917,"vote_count":9143},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1310.8526831184201,"type":"confidence_interval","upper":1328.9189950862572}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1319.8858391023387,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b32d3d46861397e196ff","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7f33c292a76091100ba734a2","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1315.5218357271092,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1004,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=3","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1004,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1315.5218357271092,"rating_lower":1308.6686146295046,"rating_upper":1322.3750568247137,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.226250912711045,"vote_count":21122},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1308.6686146295046,"type":"confidence_interval","upper":1322.3750568247137}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1315.5218357271092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7a8715a78fa2bfc15b18","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-380c9fe8040247213057bd98","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1338.363607600836,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1413,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=412","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1413,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1338.363607600836,"rating_lower":1328.9237070212785,"rating_upper":1347.8035081803937,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.19736514468987,"vote_count":5798},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1328.9237070212785,"type":"confidence_interval","upper":1347.8035081803937}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1338.363607600836,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2b13866e6d1d7ac68be0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9124d2ea1240baae49da93dc","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1323.0092205490803,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1328,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=327","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1328,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1323.0092205490803,"rating_lower":1305.293747632866,"rating_upper":1340.7246934652949,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":81.69760377345912,"vote_count":1231},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1305.293747632866,"type":"confidence_interval","upper":1340.7246934652949}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1323.0092205490803,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-770c85ff1474cecde4e7","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f252ae39a7e4fc2a2ea5d501","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1309.4296577335394,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1810,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=809","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1810,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1309.4296577335394,"rating_lower":1287.283812098676,"rating_upper":1331.575503368403,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":127.66985194301363,"vote_count":782},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1287.283812098676,"type":"confidence_interval","upper":1331.575503368403}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1309.4296577335394,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-825c96dfb8c772ac551e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f09196ca8308e00a5a3d9a45","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1337.4369448602456,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1712,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=711","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1712,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1337.4369448602456,"rating_lower":1325.4509524929167,"rating_upper":1349.4229372275745,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":37.39829573487631,"vote_count":2913},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1325.4509524929167,"type":"confidence_interval","upper":1349.4229372275745}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1337.4369448602456,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9adcef79ea2f3243f013","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8294ce121ad98d212fb0d8d5","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1322.8715236900614,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1891,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=890","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1891,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1322.8715236900614,"rating_lower":1316.1009827784444,"rating_upper":1329.6420646016784,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.933025023966547,"vote_count":15271},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1316.1009827784444,"type":"confidence_interval","upper":1329.6420646016784}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1322.8715236900614,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2fc88f97565c65cbc43b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7a6c3bcdb8d8f4b91ada46c8","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1358.969335041679,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1191,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=190","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1191,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1358.969335041679,"rating_lower":1337.7415628528302,"rating_upper":1380.1971072305278,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":117.30395486062832,"vote_count":1139},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1337.7415628528302,"type":"confidence_interval","upper":1380.1971072305278}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1358.969335041679,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-82a5b702badcd16e0e58","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-00ce8e509765e4db55ad5050","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1263.3991203690257,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3d4a76f07af37f8606cff1dc7b9772054e87d2","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1672,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=671","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1672,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1263.3991203690257,"rating_lower":1221.4099557274528,"rating_upper":1305.3882850105986,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":458.96364625835946,"vote_count":139},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.4099557274528,"type":"confidence_interval","upper":1305.3882850105986}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1263.3991203690257,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3849425ab7e6b4623093","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-228c043b7f183ffd7696eb24","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1327.2926742680615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1720,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=719","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1720,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1327.2926742680615,"rating_lower":1311.843551924762,"rating_upper":1342.7417966113608,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":62.1314433705436,"vote_count":1582},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1311.843551924762,"type":"confidence_interval","upper":1342.7417966113608}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1327.2926742680615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1384f52029a6786ad695","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1320625131a7853a3b559217","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1291.734806271278,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1530,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=529","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1530,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1291.734806271278,"rating_lower":1281.432445998864,"rating_upper":1302.0371665436921,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.62977091172838,"vote_count":5464},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1281.432445998864,"type":"confidence_interval","upper":1302.0371665436921}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1291.734806271278,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-272e6e77454b95c9e05c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a981d5d6479d425dd96e1fc4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1346.5824304077719,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1199,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=198","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1199,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1346.5824304077719,"rating_lower":1320.862804448859,"rating_upper":1372.3020563666846,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":172.1999871254261,"vote_count":672},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1320.862804448859,"type":"confidence_interval","upper":1372.3020563666846}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1346.5824304077719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-203eb4806718eb164cb3","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7b330ac1088f7a19e5c30573","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1287.2702102794583,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1022,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=21","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1022,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1287.2702102794583,"rating_lower":1279.6276540078336,"rating_upper":1294.9127665510828,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.204813871828705,"vote_count":13354},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1279.6276540078336,"type":"confidence_interval","upper":1294.9127665510828}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.2702102794583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fbea76a9008fb65d931e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6f02408b44df2c46d07be672","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1298.9715749528548,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1906,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=905","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1906,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1298.9715749528548,"rating_lower":1291.0799765073891,"rating_upper":1306.8631733983202,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.211894733579495,"vote_count":9504},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1291.0799765073891,"type":"confidence_interval","upper":1306.8631733983202}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1298.9715749528548,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c9e681b2d23a00d75f81","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2e929dfdc3eb0e8a6154d0fe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1305.244908350073,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1431,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=430","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1431,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1305.244908350073,"rating_lower":1294.3245528249956,"rating_upper":1316.1652638751505,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.04397843643667,"vote_count":3578},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1294.3245528249956,"type":"confidence_interval","upper":1316.1652638751505}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1305.244908350073,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f79a49e450b475f50067","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-800b9c3e5c30edeedbac1149","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1274.6554218727329,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1346,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=345","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1346,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1274.6554218727329,"rating_lower":1252.8500828028853,"rating_upper":1296.4607609425807,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":123.7740228763169,"vote_count":810},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1252.8500828028853,"type":"confidence_interval","upper":1296.4607609425807}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1274.6554218727329,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8ccfaf1a2ceebce8326b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-904825acd47ce770e2cddbe0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1266.5761668794835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1837,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=836","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1837,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1266.5761668794835,"rating_lower":1238.8948710566074,"rating_upper":1294.2574627023596,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":199.46956981700015,"vote_count":505},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.8948710566074,"type":"confidence_interval","upper":1294.2574627023596}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1266.5761668794835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f4612ad6fd8cf96b400c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2ac4ddf1d110ef1d2bd85cc7","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1340.212438247605,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1710,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=709","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1710,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1340.212438247605,"rating_lower":1323.9758630888327,"rating_upper":1356.449013406377,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":68.62662993191205,"vote_count":1446},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1323.9758630888327,"type":"confidence_interval","upper":1356.449013406377}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1340.212438247605,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2320331c0ee8836557ad","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-104db0ff1febd3de578f58e9","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1307.675216355538,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1333,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=332","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1333,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1307.675216355538,"rating_lower":1285.7227660506815,"rating_upper":1329.6276666603944,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":125.44975668907851,"vote_count":807},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1285.7227660506815,"type":"confidence_interval","upper":1329.6276666603944}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1307.675216355538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a7a41c4b53444130a65f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-519e1749e42eccfbd01e46a8","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1291.6511377416784,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1816,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=815","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1816,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1291.6511377416784,"rating_lower":1263.8004671151593,"rating_upper":1319.5018083681975,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":201.91804482410006,"vote_count":472},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1263.8004671151593,"type":"confidence_interval","upper":1319.5018083681975}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1291.6511377416784,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-516e42c5ef0bc19a4c0d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-27b89d99f429397f5ce8156b","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1317.0928207349784,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1423,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=422","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1423,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1317.0928207349784,"rating_lower":1306.0256415186514,"rating_upper":1328.1599999513057,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.88435995890057,"vote_count":3561},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1306.0256415186514,"type":"confidence_interval","upper":1328.1599999513057}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1317.0928207349784,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-84db9c8dbb7595a49d33","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0f11a15cf5704b25f9f38f70","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1305.235014637116,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1899,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=898","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1899,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1305.235014637116,"rating_lower":1297.3225835294404,"rating_upper":1313.1474457447914,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.297601759114812,"vote_count":9226},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1297.3225835294404,"type":"confidence_interval","upper":1313.1474457447914}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1305.235014637116,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0cd4fe9292d0192d2e99","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ecc5097f7115541bfb4ab6a9","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1293.284087685157,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1018,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=17","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1018,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1293.284087685157,"rating_lower":1285.625730045726,"rating_upper":1300.9424453245877,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.267752296985714,"vote_count":12894},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1285.625730045726,"type":"confidence_interval","upper":1300.9424453245877}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1293.284087685157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-483ce0f4f30e373a1cc2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b690e32c00b76108416891d0","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1291.0826155661205,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1531,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=530","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1531,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1291.0826155661205,"rating_lower":1280.8063187279222,"rating_upper":1301.3589124043185,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.49014935104275,"vote_count":5350},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1280.8063187279222,"type":"confidence_interval","upper":1301.3589124043185}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1291.0826155661205,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-919084fd880e7ac2b4f9","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-53a6ee166fe47b24b1bda8cb","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1344.3851109643263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63e8a2f891ca310aa14bb27132d2dc4452e538f8","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1200,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=199","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1200,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1344.3851109643263,"rating_lower":1317.056293263881,"rating_upper":1371.7139286647716,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":194.42204427151015,"vote_count":601},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1317.056293263881,"type":"confidence_interval","upper":1371.7139286647716}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1344.3851109643263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47f993e6dff7d653d5fb","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-eb9a7f1e4d244b382284cac0","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1409.7286304046831,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1180,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=179","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1180,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1409.7286304046831,"rating_lower":1375.3773178354243,"rating_upper":1444.079942973942,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":307.17827010773607,"vote_count":337},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1375.3773178354243,"type":"confidence_interval","upper":1444.079942973942}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1409.7286304046831,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9669fabf8052b3df45ef","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7d9a2a5ffe328c137b61b5b3","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1348.1123538602462,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1411,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=410","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1411,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1348.1123538602462,"rating_lower":1333.3531161849194,"rating_upper":1362.871591535573,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":56.70634696988063,"vote_count":1779},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1333.3531161849194,"type":"confidence_interval","upper":1362.871591535573}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1348.1123538602462,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4efe4b3eaa011fbc4d1f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1ede6cff42245f6a3ef0a412","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1342.11414493275,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1708,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=707","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1708,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1342.11414493275,"rating_lower":1317.3202874937278,"rating_upper":1366.908002371772,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":160.02654080135906,"vote_count":588},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1317.3202874937278,"type":"confidence_interval","upper":1366.908002371772}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1342.11414493275,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-08604b7070d35146f385","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0b3f953b891a6afd4e3b1ff9","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1333.0919622997974,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1886,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=885","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1886,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1333.0919622997974,"rating_lower":1323.7304607098863,"rating_upper":1342.4534638897082,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.813653902990307,"vote_count":4815},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1323.7304607098863,"type":"confidence_interval","upper":1342.4534638897082}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1333.0919622997974,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e3152df5bab74e62acaa","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-031539d43962f0fc3df65acf","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1322.8117958326172,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1002,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=1","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1002,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1322.8117958326172,"rating_lower":1314.2534662083392,"rating_upper":1331.370125456895,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.066976733739242,"vote_count":7049},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1314.2534662083392,"type":"confidence_interval","upper":1331.370125456895}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1322.8117958326172,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fe59eb3dfcfa03712ca1","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e05754b1465433f3c945018c","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1318.806514049852,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1517,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=516","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1517,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1318.806514049852,"rating_lower":1305.690962343518,"rating_upper":1331.9220657561857,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.779263449304786,"vote_count":2526},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1305.690962343518,"type":"confidence_interval","upper":1331.9220657561857}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1318.806514049852,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3cfd49893f52a3e30a27","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d276a39cb10f7a506ce2c6ed","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1294.7404400117455,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1337,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=336","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1337,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1294.7404400117455,"rating_lower":1269.2686166626424,"rating_upper":1320.2122633608485,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":168.8977586412534,"vote_count":603},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1269.2686166626424,"type":"confidence_interval","upper":1320.2122633608485}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1294.7404400117455,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-593ad29db7d25b738e33","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0d297e3a09334b7f4ca7043a","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1288.9859217042076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e77a46596e808a5417cbffba921b8c466eb513d8","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1818,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=817","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1818,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1288.9859217042076,"rating_lower":1251.982728701513,"rating_upper":1325.989114706902,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":356.4365404670004,"vote_count":264},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1251.982728701513,"type":"confidence_interval","upper":1325.989114706902}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1288.9859217042076,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e9798b967445be599617","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bf7f2059fbebee177049669e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1296.829101074667,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1525,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=524","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1525,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1296.829101074667,"rating_lower":1288.2242118993024,"rating_upper":1305.4339902500317,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.274999727038576,"vote_count":11061},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1288.2242118993024,"type":"confidence_interval","upper":1305.4339902500317}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1296.829101074667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8f493168074557ceaf66","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5426afa5e14c8d9cd0f07c80","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1305.0716357393028,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1432,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=431","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1432,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1305.0716357393028,"rating_lower":1296.3280870097371,"rating_upper":1313.8151844688684,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.901201068315423,"vote_count":7068},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1296.3280870097371,"type":"confidence_interval","upper":1313.8151844688684}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1305.0716357393028,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6a9302180ead9b6411e1","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1eb75aa20595009227acc3b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1310.6321671807964,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1729,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=728","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1729,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1310.6321671807964,"rating_lower":1299.6722861266474,"rating_upper":1321.5920482349452,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.269108515235306,"vote_count":3549},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1299.6722861266474,"type":"confidence_interval","upper":1321.5920482349452}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1310.6321671807964,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e3e25083f5f1c3a32fe3","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e4f5fa13ae6d303b5d5e79fe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1295.4393376689159,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1908,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=907","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1908,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1295.4393376689159,"rating_lower":1289.138231064535,"rating_upper":1301.740444273297,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.3356423413647,"vote_count":18548},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1289.138231064535,"type":"confidence_interval","upper":1301.740444273297}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1295.4393376689159,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c0274c1854ba9e16434f","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6fbabf9fba70a378f23231b5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1281.4891028508862,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1024,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=23","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1024,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1281.4891028508862,"rating_lower":1275.0683543714506,"rating_upper":1287.9098513303215,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.731863326006843,"vote_count":25559},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1275.0683543714506,"type":"confidence_interval","upper":1287.9098513303215}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1281.4891028508862,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-52beb989c7549af1c1ef","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5ae72831e758f4a674f1eb12","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1265.842434488888,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1353,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=352","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1353,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1265.842434488888,"rating_lower":1249.5540516337237,"rating_upper":1282.1308173440527,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":69.06527660980525,"vote_count":1474},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.5540516337237,"type":"confidence_interval","upper":1282.1308173440527}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1265.842434488888,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-598e355428b7ddcafbaa","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ee8b6076380004989293815f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1309.6732366965077,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1215,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=214","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1215,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1309.6732366965077,"rating_lower":1289.8061760474166,"rating_upper":1329.540297345599,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":102.74745018960007,"vote_count":1390},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1289.8061760474166,"type":"confidence_interval","upper":1329.540297345599}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1309.6732366965077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8b2a6d34f4a38a865758","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a6fd86de0ea2edd0d137d3b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1188.9836903273174,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1697,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=696","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1697,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1188.9836903273174,"rating_lower":1139.296199940722,"rating_upper":1238.6711807139127,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":642.6846716716612,"vote_count":140},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1139.296199940722,"type":"confidence_interval","upper":1238.6711807139127}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1188.9836903273174,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-595ae87cf6975bfce9bb","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d79395e4ee36c94599bf65d2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1239.3029902712422,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d941cfae876ebff798b2f46965f718e7247f5cbb","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1847,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=846","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1847,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1239.3029902712422,"rating_lower":1218.92193988579,"rating_upper":1259.6840406566946,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":108.1326741228192,"vote_count":941},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1218.92193988579,"type":"confidence_interval","upper":1259.6840406566946}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1239.3029902712422,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-12a5a2df7252c853eeb7","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a5990b69aa82910ea5c41afd","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1322.0258604166127,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1722,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=721","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1722,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1322.0258604166127,"rating_lower":1299.4838930417718,"rating_upper":1344.5678277914537,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":132.27794878107682,"vote_count":741},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1299.4838930417718,"type":"confidence_interval","upper":1344.5678277914537}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1322.0258604166127,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fc39ca41286468fae4d5","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ecbdb68c8d03a2e4219c0e2a","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1310.9250368004123,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1427,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=426","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1427,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1310.9250368004123,"rating_lower":1297.22587334882,"rating_upper":1324.6242002520048,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.85307588421271,"vote_count":2065},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1297.22587334882,"type":"confidence_interval","upper":1324.6242002520048}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1310.9250368004123,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1999d099941e9b15d81b","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d2bb56a9cf8b8c98cfaa7bad","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1296.412961143429,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1907,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=906","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1907,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1296.412961143429,"rating_lower":1287.340688271735,"rating_upper":1305.4852340151226,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.425749669652937,"vote_count":5501},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1287.340688271735,"type":"confidence_interval","upper":1305.4852340151226}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1296.412961143429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3d7d3c69fe46af5a16af","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d0dfffae6a102685ae6af6f6","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1279.2597667873204,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1827,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=826","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1827,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1279.2597667873204,"rating_lower":1245.6328553442759,"rating_upper":1312.8866782303646,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":294.3593113915037,"vote_count":323},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1245.6328553442759,"type":"confidence_interval","upper":1312.8866782303646}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1279.2597667873204,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-043713ff8cda1dbc6bbb","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3330fed95b1db2401c5ee7e3","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1272.5797654897196,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1347,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=346","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1347,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1272.5797654897196,"rating_lower":1245.7177887042433,"rating_upper":1299.4417422751962,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":187.83640031135104,"vote_count":540},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1245.7177887042433,"type":"confidence_interval","upper":1299.4417422751962}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1272.5797654897196,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-46bc8154e273533c26d3","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2dc5d7db938426ff5b9fed05","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1338.9932788275976,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1203,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=202","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1203,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1338.9932788275976,"rating_lower":1307.0197334607092,"rating_upper":1370.966824194486,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":266.12483721581395,"vote_count":389},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1307.0197334607092,"type":"confidence_interval","upper":1370.966824194486}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1338.9932788275976,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f6977cddef79a783d077","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4ed09589ed46364d26112663","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1281.1515471513026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1025,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=24","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1025,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1281.1515471513026,"rating_lower":1272.6582643696768,"rating_upper":1289.6448299329284,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.7782443534373,"vote_count":7862},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1272.6582643696768,"type":"confidence_interval","upper":1289.6448299329284}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1281.1515471513026,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-737561b27569d556a04b","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-60d33b0c20e81da9526fdd04","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1278.225200016402,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9293d4671208cf5a4855341f74010e74dcaeab41","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1538,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=537","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1538,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1278.225200016402,"rating_lower":1266.3770794177021,"rating_upper":1290.0733206151015,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.54287818083774,"vote_count":3216},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1266.3770794177021,"type":"confidence_interval","upper":1290.0733206151015}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1278.225200016402,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5b7e171fcbdccbe45ce3","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-074ebd8e0c807b7b377cdcb3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1277.975981927435,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1153,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=152","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1153,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1277.975981927435,"rating_lower":1253.5872898907644,"rating_upper":1302.3646739641056,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":154.83917100849774,"vote_count":582},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1253.5872898907644,"type":"confidence_interval","upper":1302.3646739641056}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1277.975981927435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7f8cb7363c33c0cc4433","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f315ab53cf8b729fd54625cc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1253.054839283779,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1299,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=298","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1299,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1253.054839283779,"rating_lower":1238.269469149889,"rating_upper":1267.84020941767,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":56.907331355082775,"vote_count":1745},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.269469149889,"type":"confidence_interval","upper":1267.84020941767}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1253.054839283779,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d0aefffc76854e7b8dc4","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-efe24e630bef37f750338e57","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1250.9360691466372,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1677,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=676","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1677,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1250.9360691466372,"rating_lower":1227.1933917173865,"rating_upper":1274.6787465758878,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":146.74496273985403,"vote_count":618},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1227.1933917173865,"type":"confidence_interval","upper":1274.6787465758878}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1250.9360691466372,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-39a5646d1737ef22984b","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0a4616aa3b956f7a5c33b9b2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1285.4962061183683,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1226,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=225","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1226,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1285.4962061183683,"rating_lower":1269.8220300602648,"rating_upper":1301.1703821764718,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":63.95481679432263,"vote_count":2533},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1269.8220300602648,"type":"confidence_interval","upper":1301.1703821764718}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1285.4962061183683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2effed1aa56981686398","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-08ce87c24f33f98ad653dbfc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1245.0960097197624,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1373,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=372","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1373,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1245.0960097197624,"rating_lower":1232.1537107599784,"rating_upper":1258.0383086795466,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.604034348118574,"vote_count":2719},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1232.1537107599784,"type":"confidence_interval","upper":1258.0383086795466}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1245.0960097197624,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-626888e23c1f6115ed6e","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7782ebdcbc5d06f09a674097","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1223.3192822779315,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1856,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=855","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1856,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1223.3192822779315,"rating_lower":1208.094475104317,"rating_upper":1238.5440894515461,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":60.340293699115776,"vote_count":2140},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1208.094475104317,"type":"confidence_interval","upper":1238.5440894515461}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1223.3192822779315,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-64de6428bd98821c5344","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d828f930e1b63e1e9d7d8ebb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1241.5039645417346,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1569,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=568","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1569,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1241.5039645417346,"rating_lower":1235.1184580588722,"rating_upper":1247.8894710245972,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.614377231645468,"vote_count":26645},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1235.1184580588722,"type":"confidence_interval","upper":1247.8894710245972}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1241.5039645417346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7c8ca53fe0c5f71ed812","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7fa455adfc1bd7a88cb1f13","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1235.522601905443,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1058,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=57","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1058,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1235.522601905443,"rating_lower":1230.8062347985892,"rating_upper":1240.2389690122968,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.790539408305439,"vote_count":59372},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1230.8062347985892,"type":"confidence_interval","upper":1240.2389690122968}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1235.522601905443,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-824de6bf26c85bebc7a2","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8f443362e4717a5b7c1facfa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":61,"rawValue":1244.7952459440392,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1767,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=766","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1767,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1244.7952459440392,"rating_lower":1235.861057900029,"rating_upper":1253.7294339880493,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.778490602513237,"vote_count":4803},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1235.861057900029,"type":"confidence_interval","upper":1253.7294339880493}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1244.7952459440392,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf2818b759467b471f1a","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-04f77d0cbf256854f7958c04","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":62,"rawValue":1246.40573018332,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1471,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=470","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1471,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1246.40573018332,"rating_lower":1239.1582523968027,"rating_upper":1253.6532079698372,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.673434160767126,"vote_count":8311},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1239.1582523968027,"type":"confidence_interval","upper":1253.6532079698372}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1246.40573018332,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf813f7f554b73f55862","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-209dd98a22aa16dff3ad06e7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1238.3239190626143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3534288fde2077315a8ad11adcbd8678aba7c5b1","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1947,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=946","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1947,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1238.3239190626143,"rating_lower":1233.4176907198885,"rating_upper":1243.23014740534,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.266129008410303,"vote_count":26581},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1233.4176907198885,"type":"confidence_interval","upper":1243.23014740534}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1238.3239190626143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dfca0d7be9fc841aef2f","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f26aa28fd94cb079925a61b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1302.419804626094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1150,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=149","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1150,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1302.419804626094,"rating_lower":1279.9603329177532,"rating_upper":1324.8792763344347,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":131.31153891338957,"vote_count":793},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1279.9603329177532,"type":"confidence_interval","upper":1324.8792763344347}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1302.419804626094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a4d8e3663f9efaf0d761","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-23886caaa51b3fb81a444754","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1277.3617983725967,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1296,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=295","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1296,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1277.3617983725967,"rating_lower":1263.5289926823252,"rating_upper":1291.1946040628686,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":49.81089794169208,"vote_count":2418},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1263.5289926823252,"type":"confidence_interval","upper":1291.1946040628686}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1277.3617983725967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1ebfe5404955cf9f8288","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-79a10af4d58166d167a12736","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1283.3072105961328,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1668,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=667","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1668,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1283.3072105961328,"rating_lower":1261.4390071988346,"rating_upper":1305.1754139934312,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":124.48872736821342,"vote_count":903},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1261.4390071988346,"type":"confidence_interval","upper":1305.1754139934312}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1283.3072105961328,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-adc39c129eb32c995e09","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1b75c2112bf7e236c8298949","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1269.9078347737168,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1350,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=349","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1350,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1269.9078347737168,"rating_lower":1257.2744397939732,"rating_upper":1282.5412297534606,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.547411065407395,"vote_count":3505},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1257.2744397939732,"type":"confidence_interval","upper":1282.5412297534606}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1269.9078347737168,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f6f6f9a24cbd1a93f84a","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cbfa8b08b0836cf6bd9e2149","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1261.6808522534452,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1839,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=838","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1839,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1261.6808522534452,"rating_lower":1246.9345572922305,"rating_upper":1276.42714721466,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":56.60693638357198,"vote_count":3074},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1246.9345572922305,"type":"confidence_interval","upper":1276.42714721466}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.6808522534452,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-767a2c6edd132357920b","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dd2e6ed2adf313c5617f9cc5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1280.950443411543,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1744,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=743","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1744,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1280.950443411543,"rating_lower":1272.2221293405246,"rating_upper":1289.678757482561,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.831910239914535,"vote_count":5341},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1272.2221293405246,"type":"confidence_interval","upper":1289.678757482561}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1280.950443411543,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8e35e0941c2dd0557e9d","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0f559e37c61702543fc66a48","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1269.4436691427518,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1922,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=921","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1922,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1269.4436691427518,"rating_lower":1264.6460944307619,"rating_upper":1274.241243854742,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.9916620720060765,"vote_count":32418},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1264.6460944307619,"type":"confidence_interval","upper":1274.241243854742}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1269.4436691427518,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0c9b84e6bf7240889964","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-deb789af9cb5df83c25520a6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1262.0089168132201,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1040,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=39","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1040,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1262.0089168132201,"rating_lower":1257.3021593579729,"rating_upper":1266.7156742684674,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":5.766966867686838,"vote_count":88746},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1257.3021593579729,"type":"confidence_interval","upper":1266.7156742684674}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1262.0089168132201,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b599a78b26071527eefa","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fa452b7e6b5542892300bffa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1271.765841763226,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1450,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=449","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1450,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1271.765841763226,"rating_lower":1264.6797980767963,"rating_upper":1278.8518854496556,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.071079886499223,"vote_count":9523},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1264.6797980767963,"type":"confidence_interval","upper":1278.8518854496556}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1271.765841763226,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d5f1efc8d684f2f8288c","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d506f997a38b9fcf90ed5e9f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1261.9961975859537,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1553,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=552","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1553,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1261.9961975859537,"rating_lower":1255.667752922777,"rating_upper":1268.3246422491306,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.425521585483049,"vote_count":42439},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1255.667752922777,"type":"confidence_interval","upper":1268.3246422491306}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.9961975859537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b0937622c0641876bf3d","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-27a3abfc3bf50da46d1223fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1292.1401419513602,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4da4d58d76ce82895e736c437a33668ea06cac","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1224,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=223","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1224,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1292.1401419513602,"rating_lower":1276.7214514974082,"rating_upper":1307.5588324053124,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":61.88690974223026,"vote_count":3059},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1276.7214514974082,"type":"confidence_interval","upper":1307.5588324053124}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1292.1401419513602,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-64ceedaa3deaa805d7ed","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c3be4934cea16eb8a1fc5a96","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1335.876900901611,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1148,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=147","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1148,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1335.876900901611,"rating_lower":1285.501812308244,"rating_upper":1386.2519894949776,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":660.5952762318094,"vote_count":122},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1285.501812308244,"type":"confidence_interval","upper":1386.2519894949776}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1335.876900901611,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-89b5c236d7f09f7f448a","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cac301324492df6b72d3a845","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1327.4357508689236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1662,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=661","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1662,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1327.4357508689236,"rating_lower":1296.6758634772302,"rating_upper":1358.1956382606168,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":246.30504100488628,"vote_count":334},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1296.6758634772302,"type":"confidence_interval","upper":1358.1956382606168}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1327.4357508689236,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4498261f8c2b24cd60c8","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-79fba7f6e50be8dae49e82f0","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1325.9836234703632,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1808,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=807","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1808,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1325.9836234703632,"rating_lower":1309.7893066557226,"rating_upper":1342.1779402850036,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":68.26987072727793,"vote_count":1468},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1309.7893066557226,"type":"confidence_interval","upper":1342.1779402850036}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1325.9836234703632,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-da29becb487800f495bf","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-42387ebb6e7ff8596053867b","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1322.3070079699012,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1329,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=328","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1329,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1322.3070079699012,"rating_lower":1309.1099897562767,"rating_upper":1335.5040261835254,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":45.33727884638968,"vote_count":2311},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1309.1099897562767,"type":"confidence_interval","upper":1335.5040261835254}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1322.3070079699012,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0f6f944ef217e60f9c65","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6a8d27e4e05c0d144b426486","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1367.215332507108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1189,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=188","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1189,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1367.215332507108,"rating_lower":1350.3253544475044,"rating_upper":1384.1053105667115,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":74.2612044458527,"vote_count":2185},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1350.3253544475044,"type":"confidence_interval","upper":1384.1053105667115}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1367.215332507108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-43245ccfbfe8e8703340","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9b00f8bc7048ac0226d84d96","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1312.6024874736008,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1426,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=425","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1426,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1312.6024874736008,"rating_lower":1305.081466599846,"rating_upper":1320.1235083473557,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":14.725071287692847,"vote_count":10261},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1305.081466599846,"type":"confidence_interval","upper":1320.1235083473557}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1312.6024874736008,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9c28e6eb112d1db479a3","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5aa079b81fb0f997c741aaf2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1293.9433085543112,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1017,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=16","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1017,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1293.9433085543112,"rating_lower":1288.2049797276918,"rating_upper":1299.6816373809306,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.571852324701451,"vote_count":38674},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1288.2049797276918,"type":"confidence_interval","upper":1299.6816373809306}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1293.9433085543112,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3019df7c40f91dedadef","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0534bd22954965866dab79fc","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1316.8798165983414,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1723,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=722","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1723,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1316.8798165983414,"rating_lower":1307.7264581134607,"rating_upper":1326.0331750832222,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.810456772669564,"vote_count":5266},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1307.7264581134607,"type":"confidence_interval","upper":1326.0331750832222}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1316.8798165983414,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-10d1560c9de71da61b09","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1a297e86a1666a3826898131","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1303.4830818894375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1902,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=901","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1902,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1303.4830818894375,"rating_lower":1297.9614171416833,"rating_upper":1309.0047466371916,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.936771682243808,"vote_count":27123},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1297.9614171416833,"type":"confidence_interval","upper":1309.0047466371916}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1303.4830818894375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-49412c750fcd47a3a170","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4c890fce89d1d0879b53aaa0","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1287.058712057561,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c5493e4d5272f118590d608ded16044a54c72","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1535,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=534","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1535,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1287.058712057561,"rating_lower":1279.4427390207832,"rating_upper":1294.6746850943389,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.099223499276706,"vote_count":16103},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1279.4427390207832,"type":"confidence_interval","upper":1294.6746850943389}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.058712057561,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8b82d67962537df28b0d","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a219db01d9749425059343be","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1387.6642022682222,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1181,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=180","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1181,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1387.6642022682222,"rating_lower":1355.9863471418826,"rating_upper":1419.342057394562,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":261.22537094489826,"vote_count":427},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1355.9863471418826,"type":"confidence_interval","upper":1419.342057394562}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1387.6642022682222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c13587538f70c36aab32","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a10e3501fa8470e24939dabb","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1312.7261642378835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1007,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=6","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1007,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1312.7261642378835,"rating_lower":1304.1334111160709,"rating_upper":1321.3189173596963,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.220668412392847,"vote_count":7586},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1304.1334111160709,"type":"confidence_interval","upper":1321.3189173596963}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1312.7261642378835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-86d92aaa2eba6c7d0191","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ec994bfc0a54993b082d89f9","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1304.679722098694,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1520,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=519","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1520,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1304.679722098694,"rating_lower":1292.0992880870617,"rating_upper":1317.2601561103263,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.19979604322186,"vote_count":2874},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1292.0992880870617,"type":"confidence_interval","upper":1317.2601561103263}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1304.679722098694,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dfc11609fd4aa998c1c5","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9349f635ad04ba99c9e89ffb","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1308.561229527967,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1812,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=811","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1812,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1308.561229527967,"rating_lower":1270.0285227893157,"rating_upper":1347.0939362666184,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":386.5118846539172,"vote_count":253},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1270.0285227893157,"type":"confidence_interval","upper":1347.0939362666184}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1308.561229527967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-edcbeb83bbaa0c64eab6","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f408e89994beab7d232e23bc","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1314.1394722686966,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1894,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=893","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1894,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1314.1394722686966,"rating_lower":1304.9879591227477,"rating_upper":1323.2909854146455,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.80166357877093,"vote_count":5382},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1304.9879591227477,"type":"confidence_interval","upper":1323.2909854146455}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1314.1394722686966,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0e6eafd6b3ca58e01f7b","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6c19933d1c9a065281a383c1","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1320.13295768415,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1419,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=418","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1419,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1320.13295768415,"rating_lower":1306.1349106563016,"rating_upper":1334.1310047119985,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.00804921773192,"vote_count":2019},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1306.1349106563016,"type":"confidence_interval","upper":1334.1310047119985}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1320.13295768415,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4399eae320ce84422795","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-71ed0a759cb876e72a223299","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1290.5413180022827,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1338,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=337","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1338,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1290.5413180022827,"rating_lower":1262.6901513041958,"rating_upper":1318.3924847003696,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":201.92523795021873,"vote_count":524},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.6901513041958,"type":"confidence_interval","upper":1318.3924847003696}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1290.5413180022827,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9f0cddc2622c3ce7cdc2","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2f0f4726516108bb2f1e8591","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1308.5422569432683,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1731,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=730","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1731,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1308.5422569432683,"rating_lower":1287.8795600546298,"rating_upper":1329.204953831907,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":111.14190276146391,"vote_count":858},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1287.8795600546298,"type":"confidence_interval","upper":1329.204953831907}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1308.5422569432683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a491f25bd92a623ac66d","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-418939c58ef5c638ef014f7a","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1297.5949460560837,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5761d5d759b4158c91fa9e838c320a106674220","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1435,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=434","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1435,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1297.5949460560837,"rating_lower":1275.5094347318368,"rating_upper":1319.6804573803306,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":126.97515012416616,"vote_count":791},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1275.5094347318368,"type":"confidence_interval","upper":1319.6804573803306}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1297.5949460560837,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3fa339a8590975a22a9a","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cf38bb216a00972ca497f9e2","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1284.3176000631856,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5761d5d759b4158c91fa9e838c320a106674220","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1914,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=913","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1914,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1284.3176000631856,"rating_lower":1270.7829239736259,"rating_upper":1297.8522761527454,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":47.686950557027714,"vote_count":2151},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1270.7829239736259,"type":"confidence_interval","upper":1297.8522761527454}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1284.3176000631856,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-172244ac351ec1cd0aca","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-db6b5580db1a87c27ad05639","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1272.4891253469764,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5761d5d759b4158c91fa9e838c320a106674220","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1032,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=31","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1032,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1272.4891253469764,"rating_lower":1260.821938959856,"rating_upper":1284.156311734097,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.4352980327439,"vote_count":3165},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1260.821938959856,"type":"confidence_interval","upper":1284.156311734097}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1272.4891253469764,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d7f3f5497fec461c491f","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-905d70948e954bc03be4bb07","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1264.951758641474,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5761d5d759b4158c91fa9e838c320a106674220","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1549,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=548","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1549,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1264.951758641474,"rating_lower":1246.3170247611356,"rating_upper":1283.5864925218125,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":90.3962070139514,"vote_count":1108},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1246.3170247611356,"type":"confidence_interval","upper":1283.5864925218125}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1264.951758641474,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e09c2fb48c51a14a19f4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5dfa44709076479b5d8093da","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1282.3434805213312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5761d5d759b4158c91fa9e838c320a106674220","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1743,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=742","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1743,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1282.3434805213312,"rating_lower":1244.1445640694963,"rating_upper":1320.5423969731662,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":379.8445554677605,"vote_count":235},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1244.1445640694963,"type":"confidence_interval","upper":1320.5423969731662}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1282.3434805213312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ccef82fccbc7d72e97fb","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-af136a2a6e511d48cc652ba7","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1249.7779351419053,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5761d5d759b4158c91fa9e838c320a106674220","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1366,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=365","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1366,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1249.7779351419053,"rating_lower":1212.0704230212282,"rating_upper":1287.4854472625827,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":370.13450793001306,"vote_count":255},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1212.0704230212282,"type":"confidence_interval","upper":1287.4854472625827}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1249.7779351419053,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0664121d0e6146ca56b2","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-31091102d5183c84679a611d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1387.5118900481539,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1182,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=181","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1182,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1387.5118900481539,"rating_lower":1355.4625999190218,"rating_upper":1419.5611801772861,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":267.38722077350934,"vote_count":389},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1355.4625999190218,"type":"confidence_interval","upper":1419.5611801772861}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1387.5118900481539,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-80f6e86a314faac81707","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-11cf8960dd7c2b5a3dedabda","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1315.7027920506393,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1892,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=891","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1892,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1315.7027920506393,"rating_lower":1306.5697883511707,"rating_upper":1324.835795750108,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.713562598970093,"vote_count":5268},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1306.5697883511707,"type":"confidence_interval","upper":1324.835795750108}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1315.7027920506393,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-33eb011d1171bd1e2a2d","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-940e17d23862163fac63cb59","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1303.7316413327524,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1813,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=812","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1813,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1303.7316413327524,"rating_lower":1263.915559503453,"rating_upper":1343.5477231620516,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":412.6870666157492,"vote_count":237},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1263.915559503453,"type":"confidence_interval","upper":1343.5477231620516}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1303.7316413327524,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3c66ab9dcb24d06cc57e","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3ac0fc9cc0289bf4cbde1f60","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1308.2050350788284,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1009,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=8","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1009,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1308.2050350788284,"rating_lower":1299.7401160713823,"rating_upper":1316.6699540862746,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.653031868157715,"vote_count":7533},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1299.7401160713823,"type":"confidence_interval","upper":1316.6699540862746}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1308.2050350788284,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e0de19cebc38dff95595","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ff320d63f128bee8b1c7c448","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1300.950541557701,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1523,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=522","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1523,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1300.950541557701,"rating_lower":1288.4333170600846,"rating_upper":1313.4677660553173,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.78682511958197,"vote_count":2845},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1288.4333170600846,"type":"confidence_interval","upper":1313.4677660553173}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1300.950541557701,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b989348e9bceffd5bf6d","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d9794a6cdf5821d2944b1c05","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1330.236284451325,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1717,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=716","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1717,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1330.236284451325,"rating_lower":1308.0983254565842,"rating_upper":1352.3742434460657,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":127.57893584923754,"vote_count":737},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1308.0983254565842,"type":"confidence_interval","upper":1352.3742434460657}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1330.236284451325,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3f064510ddc2efc044ea","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0029e2ee791ec6e8ae1aa762","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1303.98819623092,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1334,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=333","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1334,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1303.98819623092,"rating_lower":1274.8526202481792,"rating_upper":1333.1237722136607,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":220.97901538683595,"vote_count":439},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1274.8526202481792,"type":"confidence_interval","upper":1333.1237722136607}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1303.98819623092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":16,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-31c2346027a8c4e81a6c","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b474d8ff8f508c518c06b73f","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1317.055515893644,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9b605ed4bb387ab982cf491fe371e50dd9ede7f","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1424,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=423","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1424,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1317.055515893644,"rating_lower":1303.0141776818157,"rating_upper":1331.0968541054724,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.32403807554716,"vote_count":1995},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1303.0141776818157,"type":"confidence_interval","upper":1331.0968541054724}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1317.055515893644,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7ee7aa44b891631ebac0","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cbe404e3c58c901ab4e923bd","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1340.9007264618651,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e41d87c32bb9f689dd1492bca6a5e80b87272e96","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1709,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=708","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1709,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1340.9007264618651,"rating_lower":1304.4690509542022,"rating_upper":1377.3324019695278,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":345.51118266466585,"vote_count":263},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1304.4690509542022,"type":"confidence_interval","upper":1377.3324019695278}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1340.9007264618651,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9fbd74d0a44dd7f16271","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-850c1b2be9e1262d766b7729","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1314.7896890894135,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e41d87c32bb9f689dd1492bca6a5e80b87272e96","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1893,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=892","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1893,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1314.7896890894135,"rating_lower":1301.3160973007402,"rating_upper":1328.263280878087,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":47.257483201395885,"vote_count":2099},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1301.3160973007402,"type":"confidence_interval","upper":1328.263280878087}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1314.7896890894135,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d0f0c682f26f104d8749","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e2d1f3667c7cef3708ce67e3","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1296.9188367436175,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e41d87c32bb9f689dd1492bca6a5e80b87272e96","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1524,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=523","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1524,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1296.9188367436175,"rating_lower":1278.5799726176635,"rating_upper":1315.2577008695712,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":87.5484947589321,"vote_count":1151},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1278.5799726176635,"type":"confidence_interval","upper":1315.2577008695712}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1296.9188367436175,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6a2423ba9bdb956fb31f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-185b526449c1e84f163183b1","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1301.9064394211337,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e41d87c32bb9f689dd1492bca6a5e80b87272e96","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1013,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=12","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1013,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1301.9064394211337,"rating_lower":1290.1523273424232,"rating_upper":1313.6605514998441,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.96528225543347,"vote_count":3119},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1290.1523273424232,"type":"confidence_interval","upper":1313.6605514998441}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1301.9064394211337,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4520cc1f97c6eb782d27","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-52f9907ecb3c0bf8ecc683c2","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1285.890335789169,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e41d87c32bb9f689dd1492bca6a5e80b87272e96","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1342,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=341","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1342,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1285.890335789169,"rating_lower":1243.5575596590731,"rating_upper":1328.223111919265,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":466.506090141297,"vote_count":217},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.5575596590731,"type":"confidence_interval","upper":1328.223111919265}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1285.890335789169,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-95ab25db64d991189bd9","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3f79066b5e8387055fd45ef8","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1310.3261723682426,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e41d87c32bb9f689dd1492bca6a5e80b87272e96","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1429,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=428","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1429,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1310.3261723682426,"rating_lower":1288.9005658827532,"rating_upper":1331.751778853732,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":119.50059461735829,"vote_count":790},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1288.9005658827532,"type":"confidence_interval","upper":1331.751778853732}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1310.3261723682426,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-22e0964147ee45ad888a","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-85fb984588a14e12527033a9","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1268.647085741865,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1157,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=156","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1157,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1268.647085741865,"rating_lower":1186.2771467802604,"rating_upper":1351.0170247034698,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":1766.205798689919,"vote_count":43},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1186.2771467802604,"type":"confidence_interval","upper":1351.0170247034698}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1268.647085741865,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-beb52489cf1fc9c69dde","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bbf14a34c6dca2b81bb1127a","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1244.1771380630535,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1679,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=678","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1679,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1244.1771380630535,"rating_lower":1202.1043209297768,"rating_upper":1286.2499551963297,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":460.79419932718895,"vote_count":190},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1202.1043209297768,"type":"confidence_interval","upper":1286.2499551963297}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1244.1771380630535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-107a67b0640d5ee8d507","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-12c4dbe80cfdc99b35254537","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1275.2358717078994,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1830,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=829","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1830,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1275.2358717078994,"rating_lower":1257.749127916359,"rating_upper":1292.7226154994398,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":79.60158437302233,"vote_count":1229},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1257.749127916359,"type":"confidence_interval","upper":1292.7226154994398}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1275.2358717078994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c53eaec104c7604623a2","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c2d5d14e076eec967c41e374","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1274.897115188257,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1539,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=538","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1539,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1274.897115188257,"rating_lower":1266.7081243023076,"rating_upper":1283.086106074206,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.45679827905631,"vote_count":13727},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1266.7081243023076,"type":"confidence_interval","upper":1283.086106074206}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1274.897115188257,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9d9ffc0130b86d58d5fe","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-48029c0af66df4e8a46bc8ee","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1330.5232576900614,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1206,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=205","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1206,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1330.5232576900614,"rating_lower":1312.389410020044,"rating_upper":1348.6571053600787,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":85.60196703084186,"vote_count":1820},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1312.389410020044,"type":"confidence_interval","upper":1348.6571053600787}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1330.5232576900614,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ddd4645fd995d2af0bc4","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c76e00033ae42f42dc1898e7","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1281.3861990362748,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1444,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=443","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1444,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1281.3861990362748,"rating_lower":1273.0881356857503,"rating_upper":1289.6842623867992,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.92492346875534,"vote_count":8717},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1273.0881356857503,"type":"confidence_interval","upper":1289.6842623867992}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1281.3861990362748,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-672343a4d1ca6a062f8c","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3418f1e4efa8c79bb7e55c97","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1261.994052083512,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1358,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=357","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1358,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1261.994052083512,"rating_lower":1247.8582324521155,"rating_upper":1276.1298717149089,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":52.017060699682894,"vote_count":1945},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1247.8582324521155,"type":"confidence_interval","upper":1276.1298717149089}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.994052083512,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8f8b167a0fa56fbcdda3","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-029204fa679bab5d7bba279d","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1263.5451694367357,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1038,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=37","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1038,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1263.5451694367357,"rating_lower":1257.6132197191378,"rating_upper":1269.4771191543339,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.160068894283635,"vote_count":32772},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1257.6132197191378,"type":"confidence_interval","upper":1269.4771191543339}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1263.5451694367357,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ff7127697819870a09fd","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-71a6176400e29d8d8678862d","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1268.5293700175055,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1923,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=922","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1923,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1268.5293700175055,"rating_lower":1262.725263726676,"rating_upper":1274.333476308335,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.769493936461982,"vote_count":23306},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.725263726676,"type":"confidence_interval","upper":1274.333476308335}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1268.5293700175055,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1141d64934379a9d1524","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f3d00622ab34cc02f96afffc","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1273.281791266762,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fadf6364cdc6e460d99887415562fe8c95bdc39","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1751,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=750","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1751,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1273.281791266762,"rating_lower":1262.947500395116,"rating_upper":1283.616082138408,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.801304870030172,"vote_count":4120},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.947500395116,"type":"confidence_interval","upper":1283.616082138408}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1273.281791266762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f5c7601b6546eb0b111f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dac809a9ec7c6d039a031fcb","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1259.6186794427995,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1673,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=672","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1673,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1259.6186794427995,"rating_lower":1229.789638788283,"rating_upper":1289.4477200973163,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":231.62337744597576,"vote_count":431},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1229.789638788283,"type":"confidence_interval","upper":1289.4477200973163}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1259.6186794427995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b22d72e592861ec211f3","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b2e4e9110ad6a2298a1b2ba1","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1201.9932948782132,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1168,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=167","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1168,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1201.9932948782132,"rating_lower":1170.567515423536,"rating_upper":1233.4190743328904,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":257.0845245076895,"vote_count":365},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1170.567515423536,"type":"confidence_interval","upper":1233.4190743328904}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1201.9932948782132,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d9bcb6363f86e28a020e","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9637763ed9058419f47f1c6f","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1188.8295591249853,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1310,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=309","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1310,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1188.8295591249853,"rating_lower":1173.403795679809,"rating_upper":1204.2553225701618,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":61.943701331553854,"vote_count":1615},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1173.403795679809,"type":"confidence_interval","upper":1204.2553225701618}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1188.8295591249853,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f27583ef0d150c1fc198","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f348e587cbebc73c69b28a74","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1210.0323470503438,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1862,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=861","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1862,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1210.0323470503438,"rating_lower":1189.672447774964,"rating_upper":1230.3922463257236,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":107.90835405303376,"vote_count":1310},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1189.672447774964,"type":"confidence_interval","upper":1230.3922463257236}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1210.0323470503438,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b7abf50a2d7f2a9c03fb","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-880facd7a70fee5b29797f71","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":62,"rawValue":1239.457444566454,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1241,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=240","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1241,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1239.457444566454,"rating_lower":1219.8026940376897,"rating_upper":1259.1121950952183,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":100.5631548793123,"vote_count":1340},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1219.8026940376897,"type":"confidence_interval","upper":1259.1121950952183}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1239.457444566454,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1fccd4a8ebbe9aa0861a","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f82aea71d7fd69a2d7763607","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1201.2199259431375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1391,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=390","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1391,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1201.2199259431375,"rating_lower":1182.5773264383909,"rating_upper":1219.8625254478839,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":90.4725346584796,"vote_count":1437},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1182.5773264383909,"type":"confidence_interval","upper":1219.8625254478839}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1201.2199259431375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d408dcc112d0515de0af","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c658d37efe83d5de50e66fbd","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1228.9053106371352,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1477,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=476","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1477,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1228.9053106371352,"rating_lower":1216.7059328085938,"rating_upper":1241.1046884656762,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.741745349911376,"vote_count":2984},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1216.7059328085938,"type":"confidence_interval","upper":1241.1046884656762}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1228.9053106371352,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-29d7bba613aca9f2534f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d2b4f539ad26eca35891894d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1229.885425732703,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1775,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=774","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1775,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1229.885425732703,"rating_lower":1214.042674617193,"rating_upper":1245.728176848213,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":65.3378767347148,"vote_count":1428},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1214.042674617193,"type":"confidence_interval","upper":1245.728176848213}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1229.885425732703,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d2aa67c0942db681a99f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ccc72f09dc82a056a6956a54","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1211.316408367025,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1069,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=68","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1069,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1211.316408367025,"rating_lower":1204.4448010759327,"rating_upper":1218.1880156581171,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.291941412627159,"vote_count":34690},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1204.4448010759327,"type":"confidence_interval","upper":1218.1880156581171}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1211.316408367025,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0e2ab97931750005d986","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ba52fe43745a2ed7b891f643","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1217.908476418474,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1953,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=952","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1953,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1217.908476418474,"rating_lower":1209.3980926720565,"rating_upper":1226.4188601648916,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.853939321468847,"vote_count":10771},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1209.3980926720565,"type":"confidence_interval","upper":1226.4188601648916}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1217.908476418474,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-85be3c03d1171be95872","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-611eed9732417e21f2decd75","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1220.0929625863241,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bee32be8d98cf11de1166315a89be1639fcee58","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1581,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=580","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1581,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1220.0929625863241,"rating_lower":1210.9681916129575,"rating_upper":1229.2177335596905,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.67443390720694,"vote_count":16724},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1210.9681916129575,"type":"confidence_interval","upper":1229.2177335596905}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1220.0929625863241,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fe872043ddb869129c89","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2c9bacc11e087e3dbd291b89","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1213.799181897436,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1164,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=163","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1164,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1213.799181897436,"rating_lower":1178.012329823622,"rating_upper":1249.58603397125,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":333.38865288724685,"vote_count":289},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1178.012329823622,"type":"confidence_interval","upper":1249.58603397125}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1213.799181897436,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f647f8c903f33d3a7de4","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f9ecba5fa592bde07c7afed0","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1223.6468529366466,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1685,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=684","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1685,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1223.6468529366466,"rating_lower":1188.1580209229778,"rating_upper":1259.1356849503156,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":327.8590911634023,"vote_count":280},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1188.1580209229778,"type":"confidence_interval","upper":1259.1356849503156}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1223.6468529366466,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-39cc1cc16f57a929fdfe","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4f6a870b82131185589d7c8b","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1181.3109549238002,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1315,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=314","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1315,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1181.3109549238002,"rating_lower":1163.143509390161,"rating_upper":1199.478400457439,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":85.9194625332727,"vote_count":1108},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1163.143509390161,"type":"confidence_interval","upper":1199.478400457439}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1181.3109549238002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8f5b02f3de284d39f010","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0b27c3eb6a4f90acf909f3ea","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1206.7717125175222,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1863,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=862","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1863,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1206.7717125175222,"rating_lower":1184.8157602801753,"rating_upper":1228.7276647548692,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":125.48978425897994,"vote_count":1066},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1184.8157602801753,"type":"confidence_interval","upper":1228.7276647548692}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1206.7717125175222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-518212b8d355cc381165","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5a30a5857d1bd5af86d5faf4","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":70,"rawValue":1216.1474920472922,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1249,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=248","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1249,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1216.1474920472922,"rating_lower":1194.2728630761085,"rating_upper":1238.0221210184761,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":124.56189561352464,"vote_count":1006},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1194.2728630761085,"type":"confidence_interval","upper":1238.0221210184761}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1216.1474920472922,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b5c69966155d347c8fe7","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d6d41dcb0afbb233424646b0","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72,"rawValue":1216.2514858098239,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1584,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=583","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1584,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1216.2514858098239,"rating_lower":1205.9271222898021,"rating_upper":1226.5758493298458,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.747917410786687,"vote_count":13991},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1205.9271222898021,"type":"confidence_interval","upper":1226.5758493298458}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1216.2514858098239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-404508ad4b23cba4e4d5","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a1193a3be4055a4d68a269d3","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72,"rawValue":1192.9181362205732,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1395,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=394","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1395,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1192.9181362205732,"rating_lower":1171.323593843309,"rating_upper":1214.5126785978373,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":121.39249234466395,"vote_count":1007},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1171.323593843309,"type":"confidence_interval","upper":1214.5126785978373}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1192.9181362205732,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ec74035eb772802165ba","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3a7708c8ec8397a204b21457","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":73,"rawValue":1202.7394433197264,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1073,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=72","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1073,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1202.7394433197264,"rating_lower":1194.8981846173388,"rating_upper":1210.5807020221139,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.005726185725006,"vote_count":29292},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1194.8981846173388,"type":"confidence_interval","upper":1210.5807020221139}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1202.7394433197264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-14de0ee6f2465f79aef7","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d02268f7769300b1a0fa04c3","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":74,"rawValue":1204.0576519797155,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1483,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=482","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1483,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1204.0576519797155,"rating_lower":1189.438304037562,"rating_upper":1218.6769999218693,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":55.636502753173936,"vote_count":1965},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1189.438304037562,"type":"confidence_interval","upper":1218.6769999218693}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1204.0576519797155,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-75db488e58c9977e322c","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-84b542ed814653a70cbe3c19","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":74,"rawValue":1199.8221347399667,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1958,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=957","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1958,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1199.8221347399667,"rating_lower":1189.673839453151,"rating_upper":1209.9704300267824,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":26.809579911049276,"vote_count":8129},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1189.673839453151,"type":"confidence_interval","upper":1209.9704300267824}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1199.8221347399667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d503b35d8c7da878138c","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-57bad14cb2ed223a79567641","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 mini.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 mini.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":77,"rawValue":1208.2126541809084,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dc4a2f34686ad874ec83975576512f4eed8f076","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1783,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=782","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1783,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1208.2126541809084,"rating_lower":1189.1301814996586,"rating_upper":1227.2951268621582,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":94.79231214688691,"vote_count":993},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1189.1301814996586,"type":"confidence_interval","upper":1227.2951268621582}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1208.2126541809084,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b37d7de950acbdfa6c0e","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b5221e5de3ed7881e6c60b2f","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1117.231591275592,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1322,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=321","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1322,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1117.231591275592,"rating_lower":1076.8326526765672,"rating_upper":1157.6305298746167,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":424.8579292678352,"vote_count":206},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1076.8326526765672,"type":"confidence_interval","upper":1157.6305298746167}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1117.231591275592,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6d081e501cbbf1b9894c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6c2eb9dc1fa202c159366fe2","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":78,"rawValue":1112.4783088166164,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1882,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=881","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1882,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1112.4783088166164,"rating_lower":1066.9146032498704,"rating_upper":1158.0420143833624,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":540.4330390812282,"vote_count":176},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1066.9146032498704,"type":"confidence_interval","upper":1158.0420143833624}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1112.4783088166164,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-49dc30093c0567af798c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-502f49172157e34f69ca4365","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":92,"rawValue":1160.112291219351,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1798,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=797","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1798,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1160.112291219351,"rating_lower":1125.5697797724772,"rating_upper":1194.6548026662247,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":310.60728560453566,"vote_count":266},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1125.5697797724772,"type":"confidence_interval","upper":1194.6548026662247}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1160.112291219351,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-60720c9228caa68eadd0","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c582ba44f4a245b4cf46732d","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":93,"rawValue":1154.7895306986513,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1977,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=976","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1977,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1154.7895306986513,"rating_lower":1139.8603594851234,"rating_upper":1169.718701912179,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.01966480082107,"vote_count":1703},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1139.8603594851234,"type":"confidence_interval","upper":1169.718701912179}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1154.7895306986513,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5dc170657d24fe69af7a","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b8f210aa0842cee4ee742758","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":94,"rawValue":1173.274753015647,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1606,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=605","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1606,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1173.274753015647,"rating_lower":1158.7207475503778,"rating_upper":1187.8287584809166,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":55.140269613723476,"vote_count":1977},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1158.7207475503778,"type":"confidence_interval","upper":1187.8287584809166}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1173.274753015647,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6201d1acb73768335fc6","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9d1000cc947329649a10b3d3","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":94,"rawValue":1162.2950013064317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1094,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=93","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1094,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1162.2950013064317,"rating_lower":1151.2820884444086,"rating_upper":1173.3079141684548,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.572445617053187,"vote_count":3984},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1151.2820884444086,"type":"confidence_interval","upper":1173.3079141684548}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1162.2950013064317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d55ec60b1fbf70e6484c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ca51bc5609aa60cec99596a0","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":95,"rawValue":1140.2694032439917,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b45ea22296adcceff26ce9187f91c1fcd51d5a11","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1504,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=503","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1504,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1140.2694032439917,"rating_lower":1114.1051285183214,"rating_upper":1166.4336779696623,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":178.20554739062894,"vote_count":525},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1114.1051285183214,"type":"confidence_interval","upper":1166.4336779696623}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1140.2694032439917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3d34e30afcf12d880e3e","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2f6d6acd56e292ddf4f4dab2","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1291.2643112763622,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1293,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=292","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1293,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1291.2643112763622,"rating_lower":1261.6585429376505,"rating_upper":1320.870079615073,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":228.1689222343578,"vote_count":411},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1261.6585429376505,"type":"confidence_interval","upper":1320.870079615073}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1291.2643112763622,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d9c87a4f2d066da762c3","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ec6a0acc7c62d59c203d18e4","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1182.2640435663832,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1171,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=170","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1171,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1182.2640435663832,"rating_lower":1133.4808972147478,"rating_upper":1231.0471899180186,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":619.5030271169418,"vote_count":122},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1133.4808972147478,"type":"confidence_interval","upper":1231.0471899180186}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1182.2640435663832,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-31772b1f8e95ed0c539a","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2f45cc22bab7d48619c25557","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1254.939859587545,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1362,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=361","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1362,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1254.939859587545,"rating_lower":1231.739105771786,"rating_upper":1278.140613403304,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":140.1225426964739,"vote_count":697},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1231.739105771786,"type":"confidence_interval","upper":1278.140613403304}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1254.939859587545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0ff8b85a01b235c3dab9","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7956a3b258962bda18c07dd0","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1245.0680842058491,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1846,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=845","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1846,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1245.0680842058491,"rating_lower":1211.4619208153404,"rating_upper":1278.6742475963579,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":293.9961797183875,"vote_count":322},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1211.4619208153404,"type":"confidence_interval","upper":1278.6742475963579}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1245.0680842058491,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0dd55f0e283a9138488d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4fa33b15b6ec678486726e0d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1183.2305966439017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1701,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=700","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1701,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1183.2305966439017,"rating_lower":1139.8359884854158,"rating_upper":1226.6252048023878,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":490.20231769353643,"vote_count":140},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1139.8359884854158,"type":"confidence_interval","upper":1226.6252048023878}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1183.2305966439017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cac55e68a317f8144ebd","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a22a425768287c038840504d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1261.8111902855208,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1237,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=236","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1237,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1261.8111902855208,"rating_lower":1236.5287256921983,"rating_upper":1287.0936548788434,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":166.39590471963254,"vote_count":561},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1236.5287256921983,"type":"confidence_interval","upper":1287.0936548788434}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.8111902855208,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6a2d852e38ebfbad4527","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7c19b93fb044010b7b8c5323","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1242.8759948131935,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1943,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=942","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1943,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1242.8759948131935,"rating_lower":1234.2346630704747,"rating_upper":1251.5173265559126,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.438608552943034,"vote_count":6405},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1234.2346630704747,"type":"confidence_interval","upper":1251.5173265559126}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1242.8759948131935,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fda9057c7e2288fc5fe0","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4554be4d1ce594ebd6a887e4","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":60,"rawValue":1232.9799064050783,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1060,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=59","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1060,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1232.9799064050783,"rating_lower":1225.2159128629332,"rating_upper":1240.7438999472233,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.691850033049985,"vote_count":10027},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1225.2159128629332,"type":"confidence_interval","upper":1240.7438999472233}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1232.9799064050783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-babe950e1aeade730884","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4f71df717dd635712085dcf6","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":62,"rawValue":1242.5364839891727,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1768,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=767","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1768,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1242.5364839891727,"rating_lower":1226.8439875693625,"rating_upper":1258.2289804089824,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.10440808558538,"vote_count":1365},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1226.8439875693625,"type":"confidence_interval","upper":1258.2289804089824}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1242.5364839891727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-81bb031180fb325331f6","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ff4100cd89c55e30eac9a872","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":64,"rawValue":1245.8305564779262,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1473,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=472","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1473,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1245.8305564779262,"rating_lower":1233.2409329050897,"rating_upper":1258.4201800507626,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.26000800838389,"vote_count":2330},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1233.2409329050897,"type":"confidence_interval","upper":1258.4201800507626}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1245.8305564779262,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-51e3d063c337f33a95d7","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ad28b39166e34cc91dbb29b5","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":64,"rawValue":1231.4428551515907,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1576,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=575","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1576,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1231.4428551515907,"rating_lower":1220.7157686093276,"rating_upper":1242.1699416938538,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.954866381832094,"vote_count":4093},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1220.7157686093276,"type":"confidence_interval","upper":1242.1699416938538}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1231.4428551515907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4ffae188acf8e7192821","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0071934ea056f5428cebd381","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-01-09","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1282.407597850386,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1295,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=294","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1295,"metadata":{"category":"creative_writing","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1282.407597850386,"rating_lower":1252.1631811018065,"rating_upper":1312.6520145989657,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":238.1191071303689,"vote_count":385},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1252.1631811018065,"type":"confidence_interval","upper":1312.6520145989657}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1282.407597850386,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-79ff6ae7c1bc79860ffa","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-059255f894d39ee4963ea32b","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1243.2963597355006,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1159,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=158","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1159,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1243.2963597355006,"rating_lower":1187.2012448741534,"rating_upper":1299.3914745968477,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":819.1320168152034,"vote_count":98},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1187.2012448741534,"type":"confidence_interval","upper":1299.3914745968477}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1243.2963597355006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8772d78e3c6005a51d56","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4cde87c15da2c6431d907373","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1242.6841746921289,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1680,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=679","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1680,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1242.6841746921289,"rating_lower":1192.9646342423976,"rating_upper":1292.4037151418602,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":643.5140445123362,"vote_count":107},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1192.9646342423976,"type":"confidence_interval","upper":1292.4037151418602}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1242.6841746921289,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a86ce0e3ee74878b4028","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-536009653a355aa11a11fd27","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1253.3957374079444,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1364,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=363","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1364,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1253.3957374079444,"rating_lower":1229.8975035350393,"rating_upper":1276.8939712808497,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":143.73888174232144,"vote_count":633},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1229.8975035350393,"type":"confidence_interval","upper":1276.8939712808497}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1253.3957374079444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b3ead36ba54a219069a5","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b2a4b99cef017c49a836953c","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1272.0415127715962,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1752,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=751","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1752,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1272.0415127715962,"rating_lower":1255.3199962072845,"rating_upper":1288.7630293359077,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":72.78722205852573,"vote_count":1169},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1255.3199962072845,"type":"confidence_interval","upper":1288.7630293359077}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1272.0415127715962,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3a601501e52193856a08","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ece0b8a396103edafe3719ee","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1234.2638406611563,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1851,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=850","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1851,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1234.2638406611563,"rating_lower":1201.511868276278,"rating_upper":1267.015813046035,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":279.2407117111868,"vote_count":290},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1201.511868276278,"type":"confidence_interval","upper":1267.015813046035}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1234.2638406611563,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c923e47974ec6986319e","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bea2852c52f079998214a9c3","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1248.6543411367695,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1048,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=47","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1048,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1248.6543411367695,"rating_lower":1240.6238712773315,"rating_upper":1256.6848109962077,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.787488600929546,"vote_count":9184},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1240.6238712773315,"type":"confidence_interval","upper":1256.6848109962077}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1248.6543411367695,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c73f9bb1f083e39105e5","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-728b381ce09977deff6d1c5c","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1252.3508828972876,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1561,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=560","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1561,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1252.3508828972876,"rating_lower":1241.2827589934484,"rating_upper":1263.419006801127,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.889803449357675,"vote_count":3852},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1241.2827589934484,"type":"confidence_interval","upper":1263.419006801127}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1252.3508828972876,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3f272e52f3bfeb50df01","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5c91957f843cbb938913cb06","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1262.3312416657832,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1458,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=457","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1458,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1262.3312416657832,"rating_lower":1249.4960559807555,"rating_upper":1275.1664273508106,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":42.88526813867205,"vote_count":2160},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.4960559807555,"type":"confidence_interval","upper":1275.1664273508106}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1262.3312416657832,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-99c105cc24f023644b06","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c5409ab61bfb41980f3b18bc","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":51,"rawValue":1279.0259454710808,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1230,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=229","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1230,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1279.0259454710808,"rating_lower":1251.0463856585225,"rating_upper":1307.0055052836392,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":203.79126885006718,"vote_count":462},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1251.0463856585225,"type":"confidence_interval","upper":1307.0055052836392}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1279.0259454710808,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":22,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-088a909a62b6d52ed59d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dd45dc96fcbf9b0d0a166807","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1252.1692312332254,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0da8664b049fbf3cbec85688347974c5617a24e8","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1936,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=935","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1936,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1252.1692312332254,"rating_lower":1243.2634907312136,"rating_upper":1261.0749717352373,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.646378782434706,"vote_count":5710},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.2634907312136,"type":"confidence_interval","upper":1261.0749717352373}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1252.1692312332254,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5151c013376d441011dd","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4188bfce9c4362c522d3d5e2","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1166.9589829971826,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1173,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=172","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1173,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1166.9589829971826,"rating_lower":1109.8946491827974,"rating_upper":1224.023316811568,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":847.6827022425821,"vote_count":94},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1109.8946491827974,"type":"confidence_interval","upper":1224.023316811568}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1166.9589829971826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-054327d5f3a196c6301e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7ddef859f7fe169f15c398cd","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1207.3004207354725,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1691,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=690","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1691,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1207.3004207354725,"rating_lower":1166.2819770561443,"rating_upper":1248.3188644148008,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":437.98796249238734,"vote_count":159},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1166.2819770561443,"type":"confidence_interval","upper":1248.3188644148008}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1207.3004207354725,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3cd04c42a2828227385e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ffb042177c4fa74306d82d3b","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1261.9939078972589,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1758,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=757","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1758,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1261.9939078972589,"rating_lower":1250.3636210403424,"rating_upper":1273.6241947541753,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.21151174275043,"vote_count":2909},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1250.3636210403424,"type":"confidence_interval","upper":1273.6241947541753}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.9939078972589,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9dff1a0f569fb040acc6","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a17ce979fcd81f50a51e8c75","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1270.1309312878475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1232,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=231","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1232,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1270.1309312878475,"rating_lower":1249.0385532967625,"rating_upper":1291.2233092789324,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":115.81235933655417,"vote_count":1057},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.0385532967625,"type":"confidence_interval","upper":1291.2233092789324}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.1309312878475,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-117415572f587132a245","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e17c57baea80fad92b2b8e11","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":61,"rawValue":1223.0603952349345,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1384,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=383","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1384,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1223.0603952349345,"rating_lower":1204.4332859508866,"rating_upper":1241.6875045189822,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":90.32224904005673,"vote_count":1070},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1204.4332859508866,"type":"confidence_interval","upper":1241.6875045189822}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1223.0603952349345,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-507535985b438d1fd400","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8df1d56c155e127b56f30579","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":62,"rawValue":1229.5374284749435,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1062,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=61","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1062,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1229.5374284749435,"rating_lower":1223.2825680925089,"rating_upper":1235.7922888573778,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.184484652806495,"vote_count":18914},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1223.2825680925089,"type":"confidence_interval","upper":1235.7922888573778}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1229.5374284749435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6749412a4479b77e6a5f","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9a7517cdd6184042720d01ef","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1246.018575021936,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1472,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=471","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1472,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1246.018575021936,"rating_lower":1236.5464036995127,"rating_upper":1255.490746344359,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.356238801259018,"vote_count":4832},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1236.5464036995127,"type":"confidence_interval","upper":1255.490746344359}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1246.018575021936,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5b20a26549aa455b9648","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-faf41e37273cc408b171b2f7","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":65,"rawValue":1229.8209084385103,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1577,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=576","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1577,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1229.8209084385103,"rating_lower":1221.0258050353393,"rating_upper":1238.6160118416813,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.13658026366461,"vote_count":7936},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.0258050353393,"type":"confidence_interval","upper":1238.6160118416813}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1229.8209084385103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-14afdc6657a3641abc3a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-056e82409d8efcc5aa9e4c56","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":65,"rawValue":1192.119362312469,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1869,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=868","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1869,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1192.119362312469,"rating_lower":1167.4673175881637,"rating_upper":1216.7714070367745,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":158.2011775878782,"vote_count":594},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1167.4673175881637,"type":"confidence_interval","upper":1216.7714070367745}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1192.119362312469,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8efce3b99d7e9843c994","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-66a569c4e65b49da08042960","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":66,"rawValue":1234.0043090485344,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1950,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=949","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1950,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1234.0043090485344,"rating_lower":1227.6017136857686,"rating_upper":1240.4069044113005,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.671265603181931,"vote_count":13064},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1227.6017136857686,"type":"confidence_interval","upper":1240.4069044113005}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1234.0043090485344,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-beba59eb119118faed54","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6cf6e000fb430f7e8a68ce8f","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1275.334826947226,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1154,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=153","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1154,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1275.334826947226,"rating_lower":1216.383087099188,"rating_upper":1334.286566795264,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":904.684338249078,"vote_count":84},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1216.383087099188,"type":"confidence_interval","upper":1334.286566795264}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1275.334826947226,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-12129670685edbb55ae5","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f399bfc585751c1d9a2a6a26","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1212.7954404620546,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1690,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=689","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1690,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1212.7954404620546,"rating_lower":1169.2631439121024,"rating_upper":1256.327737012007,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":493.31801572470545,"vote_count":149},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1169.2631439121024,"type":"confidence_interval","upper":1256.327737012007}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1212.7954404620546,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7f62d70738cbebfc26c7","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-820c6fefca64950d8dcad030","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1279.155085162679,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1747,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=746","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1747,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1279.155085162679,"rating_lower":1267.4305149879556,"rating_upper":1290.879655337402,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.78472455346271,"vote_count":2812},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1267.4305149879556,"type":"confidence_interval","upper":1290.879655337402}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1279.155085162679,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf239f5b401a23785596","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dfb1c588fba125618eb730d8","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1294.9368449319488,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1222,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=221","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1222,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1294.9368449319488,"rating_lower":1273.3016783158557,"rating_upper":1316.572011548042,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":121.84965565283142,"vote_count":980},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1273.3016783158557,"type":"confidence_interval","upper":1316.572011548042}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1294.9368449319488,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0f50138d63c574b4c332","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0998329398afc655df25e1b0","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1263.948084493556,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1456,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=455","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1456,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1263.948084493556,"rating_lower":1254.232510986941,"rating_upper":1273.6636580001712,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.57201104276975,"vote_count":4499},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1254.232510986941,"type":"confidence_interval","upper":1273.6636580001712}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1263.948084493556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d7a8d48af6640fbd49d2","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0193c479f46fb56eee36f62c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1256.455683799964,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1931,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=930","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1931,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1256.455683799964,"rating_lower":1249.9292221728329,"rating_upper":1262.982145427095,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.088157743863944,"vote_count":12221},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.9292221728329,"type":"confidence_interval","upper":1262.982145427095}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1256.455683799964,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6c0cff748cee297f7220","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5130a9b3ab0b38a7e7e3a342","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1229.8091767270755,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1853,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=852","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1853,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1229.8091767270755,"rating_lower":1205.166858129958,"rating_upper":1254.4514953241928,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":158.07636999011655,"vote_count":625},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1205.166858129958,"type":"confidence_interval","upper":1254.4514953241928}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1229.8091767270755,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ce6f5880b5f1f5dcb7cc","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-abe365af02a6ad50c06cfb0d","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1246.606547156016,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1052,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=51","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1052,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1246.606547156016,"rating_lower":1240.2709037439706,"rating_upper":1252.942190568061,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.449253608642834,"vote_count":17865},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1240.2709037439706,"type":"confidence_interval","upper":1252.942190568061}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1246.606547156016,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fdf13c2ae1fbe72ff717","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9598bde15f78b9079bcb8361","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1244.7602693452304,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1566,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=565","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1566,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1244.7602693452304,"rating_lower":1235.872663856457,"rating_upper":1253.6478748340037,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.562378776144737,"vote_count":7448},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1235.872663856457,"type":"confidence_interval","upper":1253.6478748340037}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1244.7602693452304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":20,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8c1e81626512a0131910","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4f8069979cfe2b1638467719","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1224.2536232112043,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58b7a07840f958dd52399b6bb59f7bbce3682402","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1381,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=380","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1381,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1224.2536232112043,"rating_lower":1206.1626992690458,"rating_upper":1242.344547153363,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":85.19719834503658,"vote_count":1114},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1206.1626992690458,"type":"confidence_interval","upper":1242.344547153363}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1224.2536232112043,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8340fd23d47b0e3a4aae","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2b0bd34cc069cc4f4aa06aa7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1318.6922541052932,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1422,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=421","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1422,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1318.6922541052932,"rating_lower":1309.2347001090145,"rating_upper":1328.149808101572,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.28420836133456,"vote_count":5796},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1309.2347001090145,"type":"confidence_interval","upper":1328.149808101572}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1318.6922541052932,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4341954c2bebbedb5009","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-77db52d0443573b735f23c15","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1258.7874558529852,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1674,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=673","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1674,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1258.7874558529852,"rating_lower":1208.456204368496,"rating_upper":1309.1187073374745,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":659.4460579268057,"vote_count":117},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1208.456204368496,"type":"confidence_interval","upper":1309.1187073374745}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1258.7874558529852,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-41377aa0233aaaaca7db","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-edc0e05970769a92ed8dec03","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1292.7790098586884,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1020,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=19","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1020,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1292.7790098586884,"rating_lower":1285.9840644842488,"rating_upper":1299.5739552331281,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.019205410426991,"vote_count":21227},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1285.9840644842488,"type":"confidence_interval","upper":1299.5739552331281}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1292.7790098586884,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-234030662cc6419a2598","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3439e518da4b7310e509b92c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1300.6105320537492,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1904,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=903","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1904,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1300.6105320537492,"rating_lower":1293.8314652896825,"rating_upper":1307.3895988178162,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.96309744207362,"vote_count":15322},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1293.8314652896825,"type":"confidence_interval","upper":1307.3895988178162}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1300.6105320537492,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c12356fbb626e7999bd9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-31f74f6392ca873bf0bbb514","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1288.935355180308,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1533,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=532","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1533,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1288.935355180308,"rating_lower":1279.7921144118723,"rating_upper":1298.0785959487441,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.762266798029575,"vote_count":9052},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1279.7921144118723,"type":"confidence_interval","upper":1298.0785959487441}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1288.935355180308,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-06c53cccca26e392622c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-417db59194ddbd6ab7cd686a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1311.6533468830087,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1728,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=727","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1728,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1311.6533468830087,"rating_lower":1299.6364906914098,"rating_upper":1323.6702030746076,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":37.59114426833207,"vote_count":2932},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1299.6364906914098,"type":"confidence_interval","upper":1323.6702030746076}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1311.6533468830087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-313cb1f3a7057c49cc6e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e3e4a5ea671adc17407ca8f2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1341.416188386232,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1202,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=201","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1202,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1341.416188386232,"rating_lower":1319.8575830056982,"rating_upper":1362.974793766766,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":120.98879296839773,"vote_count":1120},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1319.8575830056982,"type":"confidence_interval","upper":1362.974793766766}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1341.416188386232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9ec64d0528cfaaea2a95","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7eb41a010992c7b22ee0e1d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1266.977114552296,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1352,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=351","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1352,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1266.977114552296,"rating_lower":1249.372792815132,"rating_upper":1284.5814362894598,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":80.67563868082324,"vote_count":1240},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.372792815132,"type":"confidence_interval","upper":1284.5814362894598}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1266.977114552296,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e3ca9d216e3df1c6104c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-488e204093d5d8e0317a64ae","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1257.4700485614374,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1843,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=842","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1843,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1257.4700485614374,"rating_lower":1234.992720555327,"rating_upper":1279.9473765675477,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":131.52041916278603,"vote_count":737},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1234.992720555327,"type":"confidence_interval","upper":1279.9473765675477}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1257.4700485614374,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e4f1d499a76312cd5ca7","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3c1b3d6c9672a4f9c004beab","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1322.4451062321143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1418,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=417","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1418,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1322.4451062321143,"rating_lower":1313.1799247997826,"rating_upper":1331.7102876644462,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.346611269547367,"vote_count":6100},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1313.1799247997826,"type":"confidence_interval","upper":1331.7102876644462}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1322.4451062321143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fb10e5883a536adcfc66","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-46168d81586f916e3514ae78","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1335.0946163092026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1716,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=715","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1716,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1335.0946163092026,"rating_lower":1323.2185362750663,"rating_upper":1346.9706963433391,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.7155509301356,"vote_count":2985},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1323.2185362750663,"type":"confidence_interval","upper":1346.9706963433391}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1335.0946163092026,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-410c995cc1e5e68d94d0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1d06f13b548d5e51181b85e4","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1307.2918403062465,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1897,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=896","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1897,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1307.2918403062465,"rating_lower":1300.679982080587,"rating_upper":1313.903698531906,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.380225908115138,"vote_count":16229},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1300.679982080587,"type":"confidence_interval","upper":1313.903698531906}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1307.2918403062465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d0627e94838da1fe89d4","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-067b3ef6824687bbecbda7af","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1357.0649498594507,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1192,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=191","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1192,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1357.0649498594507,"rating_lower":1336.4669729402394,"rating_upper":1377.662926778662,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":110.44675290511063,"vote_count":1249},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1336.4669729402394,"type":"confidence_interval","upper":1377.662926778662}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1357.0649498594507,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-875d1f4f2df4852990ab","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e8baefadf41ce4099f8e3028","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1297.0944905473443,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1014,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=13","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1014,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1297.0944905473443,"rating_lower":1290.4450815914674,"rating_upper":1303.743899503221,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.50985641816802,"vote_count":22515},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1290.4450815914674,"type":"confidence_interval","upper":1303.743899503221}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1297.0944905473443,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e45f031f0317adbfe8f0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dbfe7ea6483d9484c7406687","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1295.03599253479,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1526,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=525","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1526,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1295.03599253479,"rating_lower":1286.0572337633098,"rating_upper":1304.01475130627,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.98632650755963,"vote_count":9328},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1286.0572337633098,"type":"confidence_interval","upper":1304.01475130627}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1295.03599253479,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b6c2b70cc8f543d86767","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e026dcd7669aae1d367db7b9","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1283.3210953505347,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1823,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=822","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1823,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1283.3210953505347,"rating_lower":1261.0991629916575,"rating_upper":1305.5430277094117,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":128.54863238473513,"vote_count":759},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1261.0991629916575,"type":"confidence_interval","upper":1305.5430277094117}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1283.3210953505347,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c63d8b45560f5d8a24d9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-309b171b7ebbe7edac09bdf8","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1271.5249479732902,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1348,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=347","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1348,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1271.5249479732902,"rating_lower":1253.993062158957,"rating_upper":1289.0568337876236,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":80.01309777187332,"vote_count":1266},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1253.993062158957,"type":"confidence_interval","upper":1289.0568337876236}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1271.5249479732902,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":18,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e2713d52f3fcd3f441d3","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-959caa332feab6d82c18f626","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1217.7027405834353,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627df8914875eca340c4e067579b26c965ddb622","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1688,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=687","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1688,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1217.7027405834353,"rating_lower":1168.3119554624984,"rating_upper":1267.0935257043723,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":635.0320981500895,"vote_count":134},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1168.3119554624984,"type":"confidence_interval","upper":1267.0935257043723}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1217.7027405834353,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-402b5e1dd8b7f8b5d63c","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-671a1b9eb78fc84328af5aca","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1287.2696508457348,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1738,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=737","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1738,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1287.2696508457348,"rating_lower":1276.0222404529786,"rating_upper":1298.517061238491,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.93130199954162,"vote_count":3354},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1276.0222404529786,"type":"confidence_interval","upper":1298.517061238491}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.2696508457348,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-703549f61143cdc6a8f4","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-64d0ae50f999964df2016f32","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1203.5147616249446,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1693,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=692","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1693,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1203.5147616249446,"rating_lower":1158.537755144148,"rating_upper":1248.4917681057414,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":526.6049192239154,"vote_count":159},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1158.537755144148,"type":"confidence_interval","upper":1248.4917681057414}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1203.5147616249446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b7a6db16a00ca7d24ce5","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c1e20f1168690f279cac3239","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1293.5350001134723,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1223,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=222","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1223,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1293.5350001134723,"rating_lower":1273.1281855641753,"rating_upper":1313.9418146627695,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":108.40623301909017,"vote_count":1298},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1273.1281855641753,"type":"confidence_interval","upper":1313.9418146627695}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1293.5350001134723,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5770c000ec091b08694e","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5d71f85b9e8843796500881c","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1235.6731271897654,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1848,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=847","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1848,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1235.6731271897654,"rating_lower":1213.5487518367765,"rating_upper":1257.797502542754,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":127.42242143090094,"vote_count":783},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1213.5487518367765,"type":"confidence_interval","upper":1257.797502542754}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1235.6731271897654,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fc103306432e8096984c","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bd171678a4b1e52b8550afaf","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1264.299027786259,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1455,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=454","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1455,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1264.299027786259,"rating_lower":1255.3544058157681,"rating_upper":1273.24364975675,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.82705189080617,"vote_count":6489},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1255.3544058157681,"type":"confidence_interval","upper":1273.24364975675}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1264.299027786259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d1cf2f21af0a34247b2d","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a7bfd920b2423ce06177b4a9","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":51,"rawValue":1252.9412328638289,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1935,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=934","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1935,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1252.9412328638289,"rating_lower":1246.5422377435473,"rating_upper":1259.3402279841107,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.6592678616793,"vote_count":17102},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1246.5422377435473,"type":"confidence_interval","upper":1259.3402279841107}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1252.9412328638289,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3f581a628ff8f57e488f","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d3cbc373d71980cd989b2981","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1245.177396795918,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1054,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=53","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1054,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1245.177396795918,"rating_lower":1238.6908227268907,"rating_upper":1251.663970864945,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.953037665355723,"vote_count":24061},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.6908227268907,"type":"confidence_interval","upper":1251.663970864945}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1245.177396795918,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ed6316576d7c6c205088","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-736ef97a277f93473659619c","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1242.662480228433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1568,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=567","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1568,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1242.662480228433,"rating_lower":1233.9955162420422,"rating_upper":1251.329444214824,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.554098650424635,"vote_count":10244},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1233.9955162420422,"type":"confidence_interval","upper":1251.329444214824}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1242.662480228433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-123864d1e39892118088","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-17796ac6cc61d2fa49b57c6a","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1220.2872792396493,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc627fdd084367962aee506dd21242453a3c026","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1386,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=385","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1386,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1220.2872792396493,"rating_lower":1203.7803288947848,"rating_upper":1236.7942295845135,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":70.93123274417474,"vote_count":1426},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1203.7803288947848,"type":"confidence_interval","upper":1236.7942295845135}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1220.2872792396493,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf8cb1449e7eb2ff65df","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fa31dde01cc41315a76baef0","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1142.9404226060835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1705,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=704","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1705,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1142.9404226060835,"rating_lower":1096.4232185118417,"rating_upper":1189.4576267003254,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":563.2886821768309,"vote_count":148},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1096.4232185118417,"type":"confidence_interval","upper":1189.4576267003254}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1142.9404226060835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-be70bbf24819db7125a1","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4440457b63a1e63e69c68126","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1233.9470317021821,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1242,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=241","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1242,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1233.9470317021821,"rating_lower":1213.695220923267,"rating_upper":1254.1988424810972,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":106.76564788760969,"vote_count":1374},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1213.695220923267,"type":"confidence_interval","upper":1254.1988424810972}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1233.9470317021821,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cf0299d6299f4e9ef664","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f1ecbb4207e1ad214990198a","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":71,"rawValue":1226.0625736431061,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1777,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=776","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1777,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1226.0625736431061,"rating_lower":1214.591298577869,"rating_upper":1237.5338487083432,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.25525503838455,"vote_count":3299},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1214.591298577869,"type":"confidence_interval","upper":1237.5338487083432}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1226.0625736431061,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-620f3c1b5227c74f20f4","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a92f355fc7692316d95efe7e","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72,"rawValue":1211.8757458712262,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1481,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=480","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1481,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1211.8757458712262,"rating_lower":1202.69059304023,"rating_upper":1221.0608987022226,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.96223790667003,"vote_count":6367},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1202.69059304023,"type":"confidence_interval","upper":1221.0608987022226}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1211.8757458712262,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2ebe9f35aef1458557ef","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-80fa159e76d715d373e087d0","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72,"rawValue":1206.07617573138,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1956,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=955","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1956,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1206.07617573138,"rating_lower":1199.5290349874656,"rating_upper":1212.6233164752944,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.158534796652852,"vote_count":17019},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1199.5290349874656,"type":"confidence_interval","upper":1212.6233164752944}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1206.07617573138,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4a5291c531a00929395f","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-192e69837b55c718d16bba30","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":73,"rawValue":1164.3312245550119,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1877,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=876","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1877,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1164.3312245550119,"rating_lower":1142.351307058308,"rating_upper":1186.311142051716,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":125.76388182513799,"vote_count":823},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1142.351307058308,"type":"confidence_interval","upper":1186.311142051716}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1164.3312245550119,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-63a1d046bf37e4b9d2e0","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d63e921a2e4538c10cf63bce","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":76,"rawValue":1197.7172341515723,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1076,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=75","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1076,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1197.7172341515723,"rating_lower":1191.1196521921947,"rating_upper":1204.3148161109496,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.331134796034826,"vote_count":23917},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1191.1196521921947,"type":"confidence_interval","upper":1204.3148161109496}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1197.7172341515723,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1dd151ce07cadcbb6530","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-baa71d486f6d1546f06acc9c","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":77,"rawValue":1200.6826419414947,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1589,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=588","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1589,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1200.6826419414947,"rating_lower":1191.8590288843643,"rating_upper":1209.506254998625,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.267338793935984,"vote_count":9931},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1191.8590288843643,"type":"confidence_interval","upper":1209.506254998625}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1200.6826419414947,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-49be54abe6485ddd911c","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ed9eeafa2c41051c6d4ea6dc","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of GPT-5.4 nano.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of GPT-5.4 nano.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":80,"rawValue":1165.204779654008,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a7afe9889af73a571d5fd7c5e928bfc98355254","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1403,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=402","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1403,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1165.204779654008,"rating_lower":1147.5731735863178,"rating_upper":1182.836385721698,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":80.92590524503775,"vote_count":1329},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1147.5731735863178,"type":"confidence_interval","upper":1182.836385721698}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1165.204779654008,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1629e07d3851fdd08705","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b1ce5b848c9bdc8af196cde0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1323.7962585878583,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1417,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=416","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1417,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1323.7962585878583,"rating_lower":1314.3182965869328,"rating_upper":1333.274220588784,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.384804545361778,"vote_count":5676},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1314.3182965869328,"type":"confidence_interval","upper":1333.274220588784}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1323.7962585878583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f0ef0309aaede53ac836","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-322a147d5646ddeb30b8f170","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1329.394408300886,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1718,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=717","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1718,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1329.394408300886,"rating_lower":1316.8619268459465,"rating_upper":1341.9268897558254,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.886313962887066,"vote_count":2706},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1316.8619268459465,"type":"confidence_interval","upper":1341.9268897558254}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1329.394408300886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-544cb6ca8a4bc3a83125","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-73975d060a229cc6164fe00e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1293.580997071454,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1527,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=526","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1527,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1293.580997071454,"rating_lower":1284.4422438259228,"rating_upper":1302.7197503169853,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.740910102380703,"vote_count":8635},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1284.4422438259228,"type":"confidence_interval","upper":1302.7197503169853}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1293.580997071454,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7785d9ee426b8ac7a5be","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-50d8170d6c3c30a73d1070e4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1295.3180242128233,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1015,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=14","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1015,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1295.3180242128233,"rating_lower":1288.541948393104,"rating_upper":1302.0941000325427,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.952543462717362,"vote_count":20727},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1288.541948393104,"type":"confidence_interval","upper":1302.0941000325427}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1295.3180242128233,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-888454b46ef346dfcfdf","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ff7e7ee9a6e50af50c3a1ab3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1304.866839783428,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1901,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=900","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1901,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1304.866839783428,"rating_lower":1298.0416037856482,"rating_upper":1311.692075781208,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.126603095272252,"vote_count":14827},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1298.0416037856482,"type":"confidence_interval","upper":1311.692075781208}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1304.866839783428,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-276e8c7ceeb27598d8f6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ee546d8c91a4d0fd6d49e6cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1351.9563378544017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1197,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=196","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1197,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1351.9563378544017,"rating_lower":1330.9936515384557,"rating_upper":1372.9190241703477,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":114.39253629727762,"vote_count":1209},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1330.9936515384557,"type":"confidence_interval","upper":1372.9190241703477}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1351.9563378544017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-902d1eef9a2ca066165a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-278b2b4417f8fec4c2aec886","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1276.8317204106327,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1345,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=344","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1345,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1276.8317204106327,"rating_lower":1259.458992366966,"rating_upper":1294.2044484542996,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":78.56694390509483,"vote_count":1287},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1259.458992366966,"type":"confidence_interval","upper":1294.2044484542996}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1276.8317204106327,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-89b54ee2e8ca0ed1143e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2d2dd8eb98a0d25412dd0e3c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1278.8881399064164,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1828,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=827","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1828,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1278.8881399064164,"rating_lower":1256.3498937762724,"rating_upper":1301.4263860365602,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":132.23427930203744,"vote_count":731},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1256.3498937762724,"type":"confidence_interval","upper":1301.4263860365602}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1278.8881399064164,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e20d1475ead996fb3e12","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-29c6e49f3c6750b4418b859f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1336.2833885788102,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1714,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=713","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1714,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1336.2833885788102,"rating_lower":1323.6748875496553,"rating_upper":1348.8918896079651,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.383835053963836,"vote_count":2601},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1323.6748875496553,"type":"confidence_interval","upper":1348.8918896079651}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1336.2833885788102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cc6ab7388f54944f1a18","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ff1e3756dc88f9eeaf91bf3f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1298.8071058294759,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1335,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=334","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1335,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1298.8071058294759,"rating_lower":1280.7195004276964,"rating_upper":1316.8947112312553,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":85.16594461667705,"vote_count":1207},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1280.7195004276964,"type":"confidence_interval","upper":1316.8947112312553}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1298.8071058294759,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5e99b18d970c24771151","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0177788e4a7a34f66f4ed8bc","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1291.514628420573,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1817,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=816","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1817,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1291.514628420573,"rating_lower":1268.7711993562386,"rating_upper":1314.2580574849073,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":134.65289874197882,"vote_count":701},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1268.7711993562386,"type":"confidence_interval","upper":1314.2580574849073}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1291.514628420573,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bd6d8bbf2a1c1b52a562","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0a0b6d2edfd6faa6a1b43bea","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1356.4797860041688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1193,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=192","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1193,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1356.4797860041688,"rating_lower":1334.1500296316863,"rating_upper":1378.8095423766513,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":129.799131769511,"vote_count":1000},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1334.1500296316863,"type":"confidence_interval","upper":1378.8095423766513}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1356.4797860041688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4657379ee681b7e08ab9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-895d5a0fb29ba883140450ed","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1304.9281179691434,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1900,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=899","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1900,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1304.9281179691434,"rating_lower":1298.0178435568478,"rating_upper":1311.8383923814388,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.430666234396357,"vote_count":13786},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1298.0178435568478,"type":"confidence_interval","upper":1311.8383923814388}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1304.9281179691434,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ec7bf4f48a65e55fe4d7","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4185b37e839c44a7cb784dad","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1315.5795724452862,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1425,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=424","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1425,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1315.5795724452862,"rating_lower":1305.9015933714377,"rating_upper":1325.2575515191347,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.382216060544955,"vote_count":5243},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1305.9015933714377,"type":"confidence_interval","upper":1325.2575515191347}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1315.5795724452862,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-409ce927e790a5128514","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3b1fbe982e3100157c01e108","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1293.1376867171932,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1019,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=18","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1019,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1293.1376867171932,"rating_lower":1286.2885173384645,"rating_upper":1299.986856095922,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.211798529715443,"vote_count":19406},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1286.2885173384645,"type":"confidence_interval","upper":1299.986856095922}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1293.1376867171932,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5fa63d8920399aab6a0e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dfef7eca756dab6f0dcd61aa","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1290.5116187642604,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1532,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=531","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1532,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1290.5116187642604,"rating_lower":1281.2734027921779,"rating_upper":1299.749834736343,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.216725033491386,"vote_count":8166},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1281.2734027921779,"type":"confidence_interval","upper":1299.749834736343}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1290.5116187642604,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3fd02afb37d9e46e6a1a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-74f96fe001dfad16e1d020c6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1270.842498626393,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1833,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=832","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1833,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1270.842498626393,"rating_lower":1230.8760681350686,"rating_upper":1310.8089291177173,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":415.80962878296106,"vote_count":243},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1230.8760681350686,"type":"confidence_interval","upper":1310.8089291177173}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.842498626393,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4b42cc73b61b77e9d56b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d08ab9e0051e1b90930d45f9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1275.86542240388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":18,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1448,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=447","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1448,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1275.86542240388,"rating_lower":1261.8359581827067,"rating_upper":1289.8948866250532,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.237270922409266,"vote_count":2084},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1261.8359581827067,"type":"confidence_interval","upper":1289.8948866250532}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1275.86542240388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f94616d9428899ab1645","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-30588b82f748c92b61564c07","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1250.6990774461399,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":19,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1046,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=45","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1046,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1250.6990774461399,"rating_lower":1241.8538771381413,"rating_upper":1259.5442777541382,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.3666294864714,"vote_count":7227},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1241.8538771381413,"type":"confidence_interval","upper":1259.5442777541382}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1250.6990774461399,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c476bc4b8535b69aa21d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-696e4b615ef78f66971ab6e0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1254.571315524965,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":20,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1559,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=558","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1559,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1254.571315524965,"rating_lower":1242.1961105548312,"rating_upper":1266.946520495099,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.86654685137416,"vote_count":3164},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1242.1961105548312,"type":"confidence_interval","upper":1266.946520495099}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1254.571315524965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-35da1e93215100bd0c25","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0139a1cb9f0c395dee79812e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1247.6822220510298,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":21,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1370,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=369","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1370,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1247.6822220510298,"rating_lower":1213.7898354892625,"rating_upper":1281.5746086127974,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":299.0254277005915,"vote_count":323},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1213.7898354892625,"type":"confidence_interval","upper":1281.5746086127974}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1247.6822220510298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3e3316a124002d9d0e9c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-abe70bc2f3101714142813a3","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1256.3214133691158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":22,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1932,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=931","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1932,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1256.3214133691158,"rating_lower":1246.9583044334354,"rating_upper":1265.6845223047965,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.821488667052808,"vote_count":5399},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1246.9583044334354,"type":"confidence_interval","upper":1265.6845223047965}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1256.3214133691158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-980a4da6a9df2c7e9dbf","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2a8b4c693168601984db4814","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1270.1397680387188,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":23,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1231,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=230","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1231,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1270.1397680387188,"rating_lower":1238.4913644642918,"rating_upper":1301.788171613146,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":260.7398635679833,"vote_count":440},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.4913644642918,"type":"confidence_interval","upper":1301.788171613146}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.1397680387188,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":24,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fd20c0b9d080019e7b82","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8ef1dbe2e0ce977be1dbcc04","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1252.7929433303175,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04248ad5b184455addfabaa0fa80123d83f7f668","selectionRank":24,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1764,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=763","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1764,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1252.7929433303175,"rating_lower":1234.2098074998585,"rating_upper":1271.3760791607763,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":89.89630070559612,"vote_count":1131},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1234.2098074998585,"type":"confidence_interval","upper":1271.3760791607763}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1252.7929433303175,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8cc58cf7dd057c590ded","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dc61d907026163ea04dbfba5","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1283.9878695740285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1742,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=741","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1742,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1283.9878695740285,"rating_lower":1254.9594201977643,"rating_upper":1313.0163189502928,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":219.3569975684588,"vote_count":449},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1254.9594201977643,"type":"confidence_interval","upper":1313.0163189502928}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1283.9878695740285,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-231ac6d93b7d1ea264ec","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d7549212e9298135c0408ed6","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1268.806707422959,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1451,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=450","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1451,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1268.806707422959,"rating_lower":1251.9792918405963,"rating_upper":1285.634123005322,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":73.71207876969535,"vote_count":1327},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1251.9792918405963,"type":"confidence_interval","upper":1285.634123005322}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1268.806707422959,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c106c9abd5b89744e211","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6dd7fbd10b1dfe31075f1bad","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1297.0964190188013,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1221,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=220","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1221,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1297.0964190188013,"rating_lower":1258.159760833732,"rating_upper":1336.0330772038703,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":394.65823307900695,"vote_count":250},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1258.159760833732,"type":"confidence_interval","upper":1336.0330772038703}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1297.0964190188013,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-87da5863f90a6bb23773","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fb96fe1e013e385322154850","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1258.0785392330015,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1556,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=555","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1556,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1258.0785392330015,"rating_lower":1243.100528222968,"rating_upper":1273.0565502430347,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.39990073774735,"vote_count":1846},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.100528222968,"type":"confidence_interval","upper":1273.0565502430347}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1258.0785392330015,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3811f824529ed81068d8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-333bab6e005a2122c2a400e9","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1248.502963204978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1367,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=366","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1367,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1248.502963204978,"rating_lower":1216.2609938770852,"rating_upper":1280.7449325328707,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":270.611930171085,"vote_count":363},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1216.2609938770852,"type":"confidence_interval","upper":1280.7449325328707}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1248.502963204978,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e1bc9a79d20aff59d5ce","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2e5035abd9663d21f807b769","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1253.5102872507298,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1045,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=44","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1045,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1253.5102872507298,"rating_lower":1243.853202163273,"rating_upper":1263.1673723381866,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.277051177534165,"vote_count":5155},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.853202163273,"type":"confidence_interval","upper":1263.1673723381866}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1253.5102872507298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c507f2c612c83fe490db","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d4b0f34b596284cabf11e4f6","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1256.5801748871338,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1930,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=929","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1930,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1256.5801748871338,"rating_lower":1245.8684994267337,"rating_upper":1267.291850347534,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.868858817599776,"vote_count":3563},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1245.8684994267337,"type":"confidence_interval","upper":1267.291850347534}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1256.5801748871338,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e54ab0164ee24c11a319","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-34f7638fd7582fe7f2cc565e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1214.6377754173036,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dd8a4ce156d9f24db9cdcc7bf63ff12fd6a97ba","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1861,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=860","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1861,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1214.6377754173036,"rating_lower":1171.2476849360057,"rating_upper":1258.0278658986017,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":490.10025614046316,"vote_count":191},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1171.2476849360057,"type":"confidence_interval","upper":1258.0278658986017}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1214.6377754173036,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-74f78d6f827245a4be35","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2897c0b2589b6f62232064e0","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1316.23932965299,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1724,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=723","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1724,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1316.23932965299,"rating_lower":1288.175302513333,"rating_upper":1344.3033567926466,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":205.0235746515377,"vote_count":443},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1288.175302513333,"type":"confidence_interval","upper":1344.3033567926466}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1316.23932965299,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b28d4a8d33bf30d7eb4f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6205a70e56c7a64a24fd1b67","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1297.3044350302314,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1436,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=435","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1436,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1297.3044350302314,"rating_lower":1280.3376349523103,"rating_upper":1314.2712351081525,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":74.9382769205698,"vote_count":1308},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1280.3376349523103,"type":"confidence_interval","upper":1314.2712351081525}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1297.3044350302314,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2e34a0c56168909cb3ab","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c124c36e6b991825441b0099","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1285.4465654380217,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1912,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=911","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1912,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1285.4465654380217,"rating_lower":1274.7613856117212,"rating_upper":1296.1317452643225,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.721278620852406,"vote_count":3495},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1274.7613856117212,"type":"confidence_interval","upper":1296.1317452643225}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1285.4465654380217,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-243bd7be3fa001d5c03d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-70ed4ef5f2e56f35ed746a54","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1275.4883658285985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1029,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=28","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1029,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1275.4883658285985,"rating_lower":1265.9240144673286,"rating_upper":1285.0527171898684,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.813041147034355,"vote_count":5164},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1265.9240144673286,"type":"confidence_interval","upper":1285.0527171898684}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1275.4883658285985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fc44edf729e7a835f973","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0b85dd1a4b52f159d8617ecf","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1274.090876532495,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1541,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=540","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1541,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1274.090876532495,"rating_lower":1259.0481397927508,"rating_upper":1289.133613272239,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.90572805368776,"vote_count":1817},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1259.0481397927508,"type":"confidence_interval","upper":1289.133613272239}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1274.090876532495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ba696df3477c1b100e45","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a3c2f08f971f8e85f3e58821","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1261.1758740188739,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1359,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=358","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1359,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1261.1758740188739,"rating_lower":1230.2758536480767,"rating_upper":1292.075894389671,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":248.55433924530405,"vote_count":396},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1230.2758536480767,"type":"confidence_interval","upper":1292.075894389671}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.1758740188739,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-80d78b01ff0bf078c98c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-19b0c53ead2abe1261b2a9bf","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1258.359664515613,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1842,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=841","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1842,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1258.359664515613,"rating_lower":1217.481647112126,"rating_upper":1299.2376819191002,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":434.994200077837,"vote_count":212},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1217.481647112126,"type":"confidence_interval","upper":1299.2376819191002}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1258.359664515613,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-33e2142ce552209bf433","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a8e7870fb5822b414eae77ba","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1297.6844597348656,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ddae115b24bfda27504780466879605ec76b89a","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1220,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=219","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1220,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1297.6844597348656,"rating_lower":1260.4932997696162,"rating_upper":1334.875619700115,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":360.0669548010005,"vote_count":268},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1260.4932997696162,"type":"confidence_interval","upper":1334.875619700115}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1297.6844597348656,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-acc369819c2acc402fff","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-eaedccea7035fb85151e334c","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1374.2578449041494,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1186,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=185","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1186,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1374.2578449041494,"rating_lower":1336.6145302008463,"rating_upper":1411.9011596074529,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":368.87526535970613,"vote_count":256},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1336.6145302008463,"type":"confidence_interval","upper":1411.9011596074529}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1374.2578449041494,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8ef9c2ad4539b059f32e","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e2ce6f6928e25cc08f9d844b","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1316.024938583092,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1725,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=724","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1725,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1316.024938583092,"rating_lower":1287.3386498926802,"rating_upper":1344.7112272735037,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":214.21631657135637,"vote_count":431},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1287.3386498926802,"type":"confidence_interval","upper":1344.7112272735037}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1316.024938583092,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a10b1834d40bcb6e5327","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-980634167f4eab6455ee7ae7","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1296.1252619476961,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1437,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=436","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1437,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1296.1252619476961,"rating_lower":1279.558647860955,"rating_upper":1312.691876034437,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":71.4449158794862,"vote_count":1367},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1279.558647860955,"type":"confidence_interval","upper":1312.691876034437}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1296.1252619476961,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c76081ee8a62cf0002a9","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ac42a1cf8cde9b902ac7e2de","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1268.9507176368982,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1033,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=32","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1033,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1268.9507176368982,"rating_lower":1259.3476511826927,"rating_upper":1278.5537840911036,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.006214729440554,"vote_count":5094},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1259.3476511826927,"type":"confidence_interval","upper":1278.5537840911036}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1268.9507176368982,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f6302209f240916026f7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8379f773a6e8daec48021d3c","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1279.0604106972467,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1917,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=916","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1917,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1279.0604106972467,"rating_lower":1268.3913481048298,"rating_upper":1289.7294732896632,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.63168470990857,"vote_count":3506},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1268.3913481048298,"type":"confidence_interval","upper":1289.7294732896632}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1279.0604106972467,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-04d2356e03102fc12d8e","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-833e5a38c1303c644da6f854","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1262.4022278899602,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1551,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=550","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1551,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1262.4022278899602,"rating_lower":1247.3605937762366,"rating_upper":1277.443862003684,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.8970928419995,"vote_count":1785},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1247.3605937762366,"type":"confidence_interval","upper":1277.443862003684}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1262.4022278899602,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e239758c870f57031213","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cd0a75a10c612354bf97da72","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1254.1527858213894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1363,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=362","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1363,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1254.1527858213894,"rating_lower":1221.9255845306448,"rating_upper":1286.379987112134,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":270.3640860183673,"vote_count":365},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.9255845306448,"type":"confidence_interval","upper":1286.379987112134}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1254.1527858213894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-380ac1e8ed9c7ef671d1","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3bdd82f3fb48ef042686c8ec","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1220.252084629971,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03dc4c2485993ab4b62e596a5429efdcf3f1e563","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1860,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=859","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1860,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1220.252084629971,"rating_lower":1178.772661555379,"rating_upper":1261.731507704563,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":447.88779963808344,"vote_count":200},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1178.772661555379,"type":"confidence_interval","upper":1261.731507704563}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1220.252084629971,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6098ac006f7ce71d9b63","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d740c8af82094bf80561bea0","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1319.3826244267464,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1420,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=419","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1420,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1319.3826244267464,"rating_lower":1303.6142599583002,"rating_upper":1335.1509888951925,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.72575383870137,"vote_count":1527},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1303.6142599583002,"type":"confidence_interval","upper":1335.1509888951925}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1319.3826244267464,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bc9203f00db300745444","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ffcd303a2c9b747aec39093c","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1323.7480829167566,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1721,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=720","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1721,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1323.7480829167566,"rating_lower":1295.176742450721,"rating_upper":1352.3194233827926,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":212.50299277674824,"vote_count":461},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1295.176742450721,"type":"confidence_interval","upper":1352.3194233827926}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1323.7480829167566,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bec1fc4214e7a492dd13","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6ce1c123df322be28494bf98","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1354.986672617962,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1196,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=195","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1196,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1354.986672617962,"rating_lower":1316.7364780514663,"rating_upper":1393.2368671844577,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":380.8650444183091,"vote_count":274},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1316.7364780514663,"type":"confidence_interval","upper":1393.2368671844577}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1354.986672617962,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f6e01c5ddef58c375452","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b768c875b79b3bc3217dfb8b","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1287.1952109984525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1341,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=340","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1341,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1287.1952109984525,"rating_lower":1255.4337293099525,"rating_upper":1318.9566926869522,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":262.6064123385893,"vote_count":393},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1255.4337293099525,"type":"confidence_interval","upper":1318.9566926869522}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.1952109984525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-73fe3468ffa80eee08aa","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f61d7b8e4320a5422fd46c66","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1300.110262572361,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1905,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=904","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1905,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1300.110262572361,"rating_lower":1289.8909587206413,"rating_upper":1310.329566424081,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.186070732083632,"vote_count":3965},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1289.8909587206413,"type":"confidence_interval","upper":1310.329566424081}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1300.110262572361,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-df6323f7402371f04503","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-426c848b5b4291c58e4a38f6","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1291.3320711674387,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1021,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=20","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1021,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1291.3320711674387,"rating_lower":1282.0992859123885,"rating_upper":1300.564856422489,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.19061235451936,"vote_count":5722},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1282.0992859123885,"type":"confidence_interval","upper":1300.564856422489}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1291.3320711674387,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4654459c805a1444c235","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ccf23ae6f789d15102681bdd","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1279.3052179019526,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1826,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=825","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1826,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1279.3052179019526,"rating_lower":1238.9709523541921,"rating_upper":1319.639483449713,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":423.498741809572,"vote_count":216},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.9709523541921,"type":"confidence_interval","upper":1319.639483449713}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1279.3052179019526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3477d82537ff7a683ef8","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d6e7687dbe8914215a1e8886","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1284.489086959617,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"972ab2c00de8708d0a8c24d21d6d29d2fa69ec2d","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1537,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=536","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1537,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1284.489086959617,"rating_lower":1269.970458509604,"rating_upper":1299.0077154096298,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":54.87253200112013,"vote_count":1980},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1269.970458509604,"type":"confidence_interval","upper":1299.0077154096298}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1284.489086959617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6a18dab8c7df61d6bbe0","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fff757d3843cc5c3ad299e8c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1343.6767036943584,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1201,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=200","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1201,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1343.6767036943584,"rating_lower":1301.0137863821644,"rating_upper":1386.3396210065525,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":473.81075746068467,"vote_count":173},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1301.0137863821644,"type":"confidence_interval","upper":1386.3396210065525}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1343.6767036943584,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6dbc8e7d5343fb4a4651","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9cd0fe95a63351e4e52342ec","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1251.3036843705388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1365,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=364","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1365,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1251.3036843705388,"rating_lower":1215.043945767729,"rating_upper":1287.563422973349,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":342.25764349246253,"vote_count":243},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1215.043945767729,"type":"confidence_interval","upper":1287.563422973349}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1251.3036843705388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4b5505a5deedd1600704","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-32284efd1ef4355bbbea52eb","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1258.9378101102452,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1555,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=554","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1555,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1258.9378101102452,"rating_lower":1243.2395512452515,"rating_upper":1274.6360689752391,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.15149631822113,"vote_count":1563},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.2395512452515,"type":"confidence_interval","upper":1274.6360689752391}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1258.9378101102452,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3f3b9cc3fc1d7f5383a2","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d233dee4a74012014d3d467a","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1254.80414442906,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1044,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=43","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1044,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1254.80414442906,"rating_lower":1243.8767809937021,"rating_upper":1265.7315078644178,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.08383487157062,"vote_count":3851},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.8767809937021,"type":"confidence_interval","upper":1265.7315078644178}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1254.80414442906,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9e08b5f8f41478cb3619","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-54b0aae2b0e8bdf15aab6495","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1260.7183466538447,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1928,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=927","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1928,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1260.7183466538447,"rating_lower":1248.429997818533,"rating_upper":1273.0066954891565,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.30889907887237,"vote_count":2574},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1248.429997818533,"type":"confidence_interval","upper":1273.0066954891565}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1260.7183466538447,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-000f259c2cbd44d658e3","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7ba7bb93d01af7493c4352f9","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":51,"rawValue":1257.627645091559,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":16,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1460,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=459","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1460,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1257.627645091559,"rating_lower":1238.392684240999,"rating_upper":1276.862605942119,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":96.3133372481985,"vote_count":901},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.392684240999,"type":"confidence_interval","upper":1276.862605942119}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1257.627645091559,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c194c4abd3bd4190486d","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fe47361eb17a522c4c5dfb14","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1261.959039824833,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":17,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1759,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=758","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1759,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1261.959039824833,"rating_lower":1238.444833337097,"rating_upper":1285.4732463125695,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":143.9343573772826,"vote_count":554},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.444833337097,"type":"confidence_interval","upper":1285.4732463125695}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1261.959039824833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f74fbd48fa4a7c429b57","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0dcaead144c5a13a66d57627","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1284.0378835814013,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1667,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=666","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1667,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1284.0378835814013,"rating_lower":1249.9687438463413,"rating_upper":1318.107023316461,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":302.1524729184268,"vote_count":248},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.9687438463413,"type":"confidence_interval","upper":1318.107023316461}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1284.0378835814013,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dbb146ab77c1181e2135","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-58f5aa336027e5b239b84233","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"captioning","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1253.19540869069,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1158,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=157","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1158,"metadata":{"category":"captioning","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1253.19540869069,"rating_lower":1200.6957138913078,"rating_upper":1305.6951034900724,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":717.4925159109903,"vote_count":109},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1200.6957138913078,"type":"confidence_interval","upper":1305.6951034900724}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1253.19540869069,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-90b60086b019cd9a7eb3","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-75e6dc1a9f77977795c943b6","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1284.3454563531582,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1822,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=821","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1822,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1284.3454563531582,"rating_lower":1265.1343057784532,"rating_upper":1303.5566069278632,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":96.07503910123968,"vote_count":949},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1265.1343057784532,"type":"confidence_interval","upper":1303.5566069278632}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1284.3454563531582,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2f24d08e6bb479f8cf3f","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d6263e61bbc7b8a0b7ed805a","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1296.290374421956,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1735,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=734","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1735,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1296.290374421956,"rating_lower":1285.8556286539279,"rating_upper":1306.7251201899842,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.344419223452036,"vote_count":3517},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1285.8556286539279,"type":"confidence_interval","upper":1306.7251201899842}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1296.290374421956,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-55c3c9756ee99bd70b15","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2f7b224e85824580ff10c597","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1264.6581191073087,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1355,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=354","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1355,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1264.6581191073087,"rating_lower":1249.1868781259172,"rating_upper":1280.1293600887002,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":62.30947894452184,"vote_count":1494},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.1868781259172,"type":"confidence_interval","upper":1280.1293600887002}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1264.6581191073087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fcd99e525832125c039c","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-550c155e739084ecf33fed2a","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1270.5106909568983,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1545,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=544","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1545,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1270.5106909568983,"rating_lower":1262.5433986394962,"rating_upper":1278.4779832743006,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.524385613345185,"vote_count":10958},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.5433986394962,"type":"confidence_interval","upper":1278.4779832743006}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.5106909568983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f8140e5e08104ac12dc7","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-82fd16dbfa09f17595b7b579","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1273.8052118537719,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1919,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=918","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1919,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1273.8052118537719,"rating_lower":1268.0352311800627,"rating_upper":1279.575192527481,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.666675481623594,"vote_count":18518},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1268.0352311800627,"type":"confidence_interval","upper":1279.575192527481}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1273.8052118537719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3d455aea30c79f5f16e1","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7932013c42fb025f187c195b","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1266.7468257269916,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1035,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=34","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1035,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1266.7468257269916,"rating_lower":1260.819497241429,"rating_upper":1272.6741542125542,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.145802315125009,"vote_count":26666},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1260.819497241429,"type":"confidence_interval","upper":1272.6741542125542}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1266.7468257269916,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f816c5133b09b135b589","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-324c292c4dde3ce1f3b60a4b","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1317.8025399585476,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1214,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=213","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1214,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1317.8025399585476,"rating_lower":1298.4918005091813,"rating_upper":1337.113279407914,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":97.07370962105792,"vote_count":1338},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1298.4918005091813,"type":"confidence_interval","upper":1337.113279407914}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1317.8025399585476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":17,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-310c4630d1629d786465","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-06e518aab32aff9be358df99","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1281.092411400433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87b31a6d98c238562ab974ad49e2be1d8e2fc9f3","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1445,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=444","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1445,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1281.092411400433,"rating_lower":1273.0001638130589,"rating_upper":1289.1846589878076,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.046771570893107,"vote_count":6958},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1273.0001638130589,"type":"confidence_interval","upper":1289.1846589878076}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1281.092411400433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1da90b01d270d0af4662","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0932b269597eb42b2e6b1098","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1348.6841917887145,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1198,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=197","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1198,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1348.6841917887145,"rating_lower":1325.377231769154,"rating_upper":1371.9911518082754,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":141.40835831093088,"vote_count":873},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1325.377231769154,"type":"confidence_interval","upper":1371.9911518082754}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1348.6841917887145,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8748d88dd9a99d48ca8a","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-813c50361b22d0f9a66159db","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1284.8125957988436,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1536,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=535","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1536,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1284.8125957988436,"rating_lower":1275.047835310423,"rating_upper":1294.5773562872641,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.82144202159932,"vote_count":6415},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1275.047835310423,"type":"confidence_interval","upper":1294.5773562872641}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1284.8125957988436,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0d5cfec9d62705d394d7","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-588b30a08b8582975e7c2cdd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1310.1020101239699,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1730,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=729","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1730,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1310.1020101239699,"rating_lower":1296.7539940529211,"rating_upper":1323.4500261950188,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":46.38069581096857,"vote_count":2148},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1296.7539940529211,"type":"confidence_interval","upper":1323.4500261950188}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1310.1020101239699,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6a144b0ec3d8568352ea","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bdedf861b87fb800c1b3f056","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1278.8400373640645,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1829,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=828","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1829,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1278.8400373640645,"rating_lower":1253.0963051532276,"rating_upper":1304.5837695749017,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":172.5229344053034,"vote_count":536},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1253.0963051532276,"type":"confidence_interval","upper":1304.5837695749017}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1278.8400373640645,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dcb8f19695310fab00d8","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aee7b69882e18e2f45869682","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1297.7564823939383,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1434,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=433","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1434,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1297.7564823939383,"rating_lower":1287.2536641133559,"rating_upper":1308.2593006745208,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.71544300844651,"vote_count":4045},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1287.2536641133559,"type":"confidence_interval","upper":1308.2593006745208}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1297.7564823939383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e8fa7c72b20ffd426b89","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-120fc66341c5ab6396249b53","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1290.8729564438972,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1909,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=908","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1909,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1290.8729564438972,"rating_lower":1283.5564113803096,"rating_upper":1298.1895015074851,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.935287130798375,"vote_count":11056},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1283.5564113803096,"type":"confidence_interval","upper":1298.1895015074851}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1290.8729564438972,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e1fa0d6585cc745bc89a","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-209efe70c3b8df34f810d221","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1280.7346998967805,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1026,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=25","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1026,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1280.7346998967805,"rating_lower":1273.6441101710902,"rating_upper":1287.8252896224708,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.08785672443539,"vote_count":15381},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1273.6441101710902,"type":"confidence_interval","upper":1287.8252896224708}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1280.7346998967805,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-71708c0a494eb981d50c","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e49d7f616402fcfe66876354","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1265.2797344652085,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9157e21302e6a5aa7fcb007b734054fb710095b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1354,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=353","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1354,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1265.2797344652085,"rating_lower":1244.1254002165776,"rating_upper":1286.4340687138392,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":116.49372761515976,"vote_count":814},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1244.1254002165776,"type":"confidence_interval","upper":1286.4340687138392}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1265.2797344652085,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c8840b73fe0142f85a28","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-02be7129a863131848a92fdf","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1281.2103866483121,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1229,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=228","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1229,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1281.2103866483121,"rating_lower":1253.1699781698308,"rating_upper":1309.2507951267937,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":204.67862453827485,"vote_count":550},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1253.1699781698308,"type":"confidence_interval","upper":1309.2507951267937}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1281.2103866483121,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c962a7f2859f8722fdf9","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8517a59806c74ab041364365","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1255.1672936816294,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1463,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=462","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1463,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1255.1672936816294,"rating_lower":1242.743644106989,"rating_upper":1267.59094325627,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.179284995062204,"vote_count":2670},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1242.743644106989,"type":"confidence_interval","upper":1267.59094325627}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1255.1672936816294,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-13a1bb4d4b72e92c978f","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-891295d0d8957f613d5c4518","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1253.2921022521436,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1763,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=762","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1763,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1253.2921022521436,"rating_lower":1237.6886275247448,"rating_upper":1268.8955769795425,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":63.37915748491085,"vote_count":1590},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1237.6886275247448,"type":"confidence_interval","upper":1268.8955769795425}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1253.2921022521436,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2633f21c7bbcf0833cdc","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a3d5cd67f462cf4b1fe84b43","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1241.0158810190953,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1570,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=569","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1570,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1241.0158810190953,"rating_lower":1229.9414886463785,"rating_upper":1252.0902733918122,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.925935471235682,"vote_count":4200},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1229.9414886463785,"type":"confidence_interval","upper":1252.0902733918122}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1241.0158810190953,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-223f80e4f34a033e5bca","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b96bcbdf52cbc36e72820b83","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":61,"rawValue":1238.9233609436715,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1945,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=944","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1945,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1238.9233609436715,"rating_lower":1230.7839834080642,"rating_upper":1247.0627384792788,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.245913534269757,"vote_count":7177},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1230.7839834080642,"type":"confidence_interval","upper":1247.0627384792788}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1238.9233609436715,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ef1e9afd3eb15ec46520","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-685bbeecc958684e2f078521","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1228.8261260156598,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1063,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=62","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1063,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1228.8261260156598,"rating_lower":1221.187818091422,"rating_upper":1236.4644339398978,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.187914453533963,"vote_count":10088},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.187818091422,"type":"confidence_interval","upper":1236.4644339398978}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1228.8261260156598,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6f80e5e39ca662a13d48","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-297e95708898766f2f5a6854","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1182.0435256255978,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1872,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=871","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1872,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1182.0435256255978,"rating_lower":1148.5375545498687,"rating_upper":1215.5494967013267,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":292.24577175728484,"vote_count":340},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1148.5375545498687,"type":"confidence_interval","upper":1215.5494967013267}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1182.0435256255978,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e6fdbb054cc4d9281026","canonicalModelId":"xiaomi/mimo-v2-omni@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2ba52b585da563ac672ab5da","mappingCandidates":["xiaomi/mimo-v2-omni@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Omni","modelRef":"mimo-v2-omni","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":73,"rawValue":1188.245002592475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35768900b9dd63565819e407b2bb2ed2b148d989","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1396,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=395","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1396,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1188.245002592475,"rating_lower":1162.9182571786034,"rating_upper":1213.5717480063465,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":166.97928136143884,"vote_count":585},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1162.9182571786034,"type":"confidence_interval","upper":1213.5717480063465}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1188.245002592475,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4568fc875e1211783175","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fff67674db17da2d59350a90","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1328.613451384389,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1207,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=206","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1207,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1328.613451384389,"rating_lower":1307.1046067852653,"rating_upper":1350.1222959835125,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":120.4309137708393,"vote_count":1084},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1307.1046067852653,"type":"confidence_interval","upper":1350.1222959835125}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1328.613451384389,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-206afb39d913e646728a","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5251e4990e23f46a628829fd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1195.7061799232793,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1696,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=695","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1696,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1195.7061799232793,"rating_lower":1151.6370519135476,"rating_upper":1239.7753079330105,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":505.5600317972887,"vote_count":152},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1151.6370519135476,"type":"confidence_interval","upper":1239.7753079330105}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1195.7061799232793,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-555de699250696ecbdf7","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fec608d411ee66345f54ff21","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1271.8996609080855,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1449,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=448","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1449,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1271.8996609080855,"rating_lower":1262.6093646429242,"rating_upper":1281.189957173247,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.46792396407233,"vote_count":6026},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.6093646429242,"type":"confidence_interval","upper":1281.189957173247}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1271.8996609080855,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5cb9ddfabe5dd5a9e5c8","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-79348326b53c3dbe7aba9a55","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1262.135794150942,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1552,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=551","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1552,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1262.135794150942,"rating_lower":1253.149522660987,"rating_upper":1271.1220656408968,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.02146061182659,"vote_count":9176},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1253.149522660987,"type":"confidence_interval","upper":1271.1220656408968}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1262.135794150942,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b9c997c77a5d7c5133f4","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4adb8e94d772f8ec93c1c5f6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1259.012296049082,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1929,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=928","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1929,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1259.012296049082,"rating_lower":1252.3834896396859,"rating_upper":1265.6411024584781,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.438642574153995,"vote_count":16166},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1252.3834896396859,"type":"confidence_interval","upper":1265.6411024584781}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1259.012296049082,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7eac5330735ffe075653","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-de7e809df279a6bcc240b371","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1248.5602985970718,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1049,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=48","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1049,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1248.5602985970718,"rating_lower":1241.825729705989,"rating_upper":1255.2948674881548,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.806561065919766,"vote_count":22457},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1241.825729705989,"type":"confidence_interval","upper":1255.2948674881548}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1248.5602985970718,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f9068c75478a952da6f5","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-81ade7c0248eed40878aba21","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1266.2051212157517,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1756,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=755","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1756,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1266.2051212157517,"rating_lower":1254.8196971657685,"rating_upper":1277.5905452657348,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":33.74444107004968,"vote_count":3212},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1254.8196971657685,"type":"confidence_interval","upper":1277.5905452657348}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1266.2051212157517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-613c6a1fd99e8126851c","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-52a99d0211665e00b2002b30","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1225.1121806112103,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1854,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=853","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1854,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1225.1121806112103,"rating_lower":1203.7684214046583,"rating_upper":1246.455939817762,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":118.58933762693697,"vote_count":816},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1203.7684214046583,"type":"confidence_interval","upper":1246.455939817762}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1225.1121806112103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f903a665c2c063d46928","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1d6dc4e730f3d28d425f1b6d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1223.9541695976154,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c00d1fa3faf3a12c3556b6d603fddd905068d7c","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1382,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=381","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1382,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1223.9541695976154,"rating_lower":1206.262267724279,"rating_upper":1241.646071470952,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":81.48034549000042,"vote_count":1261},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1206.262267724279,"type":"confidence_interval","upper":1241.646071470952}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1223.9541695976154,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-089cd44cb5d36cfe77ba","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-946318fec2eb4d824f373c09","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1319.7348565813618,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1211,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=210","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1211,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1319.7348565813618,"rating_lower":1292.4929958347261,"rating_upper":1346.9767173279977,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":193.1867583588989,"vote_count":573},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1292.4929958347261,"type":"confidence_interval","upper":1346.9767173279977}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1319.7348565813618,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-05d6d6748aaa5ac5903d","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c73d54049ba2dd92407d50a4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1265.6258205852478,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1548,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=547","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1548,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1265.6258205852478,"rating_lower":1255.1748989913324,"rating_upper":1276.076742179163,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.43236573922993,"vote_count":5098},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1255.1748989913324,"type":"confidence_interval","upper":1276.076742179163}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1265.6258205852478,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4cec5e20b78b10553eb7","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1abc93758580fe6201f976d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1259.189231040419,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1841,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=840","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1841,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1259.189231040419,"rating_lower":1229.7993118078782,"rating_upper":1288.5791502729596,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":224.85399240572536,"vote_count":412},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1229.7993118078782,"type":"confidence_interval","upper":1288.5791502729596}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1259.189231040419,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-520d1dfe369b473161f3","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-55779c0360d973bf3b3b75ba","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1264.213995432525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1926,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=925","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1926,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1264.213995432525,"rating_lower":1256.3077709190088,"rating_upper":1272.1202199460413,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.272043766651638,"vote_count":9092},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1256.3077709190088,"type":"confidence_interval","upper":1272.1202199460413}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1264.213995432525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8e7113925dd3e064abf6","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a2cb7ae617d76ef673e429f1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1255.1367541779075,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1043,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=42","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1043,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1255.1367541779075,"rating_lower":1247.4341613527122,"rating_upper":1262.839347003103,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.444636790361468,"vote_count":12732},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1247.4341613527122,"type":"confidence_interval","upper":1262.839347003103}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1255.1367541779075,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9f40f796968190fbdce4","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-877f300dc22d4c25c6df53e7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1265.633147559231,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1454,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=453","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1454,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1265.633147559231,"rating_lower":1254.4616110316151,"rating_upper":1276.8046840868474,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.48849830578433,"vote_count":3429},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1254.4616110316151,"type":"confidence_interval","upper":1276.8046840868474}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1265.633147559231,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-63da139ba3e91ebadb1e","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-69575293b6af5e62619176a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1270.1936130461481,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1754,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=753","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1754,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1270.1936130461481,"rating_lower":1254.3304731830958,"rating_upper":1286.0567529092,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":65.50615744184417,"vote_count":1524},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1254.3304731830958,"type":"confidence_interval","upper":1286.0567529092}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.1936130461481,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c8544fff01fc8f2bcf0e","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-648d14867e5ef8e022f10813","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1238.8376446035531,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a9a859d645d9add3adede355698d3d60d6903b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1375,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=374","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1375,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1238.8376446035531,"rating_lower":1215.8049911441046,"rating_upper":1261.8702980630014,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":138.09939144087457,"vote_count":710},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1215.8049911441046,"type":"confidence_interval","upper":1261.8702980630014}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1238.8376446035531,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7704e92e6ff0b40c61da","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-473c0c24b969a35f14c3f4bf","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1300.6831570835932,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1219,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=218","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1219,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1300.6831570835932,"rating_lower":1276.2702372633837,"rating_upper":1325.096076903803,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":155.1469589983244,"vote_count":740},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1276.2702372633837,"type":"confidence_interval","upper":1325.096076903803}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1300.6831570835932,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-28516779ec86e87ebba1","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1767733a06933dc3ad96ae48","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1250.7123541477056,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1844,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=843","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1844,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1250.7123541477056,"rating_lower":1221.9000246516841,"rating_upper":1279.524683643727,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":216.1028842780367,"vote_count":414},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.9000246516841,"type":"confidence_interval","upper":1279.524683643727}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1250.7123541477056,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-29b5bdd5746d9cb2c00c","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1f81ffef01ca3c07d546a8ad","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1246.7276612449236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1564,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=563","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1564,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1246.7276612449236,"rating_lower":1236.794096975846,"rating_upper":1256.6612255140012,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.68703810029783,"vote_count":5283},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1236.794096975846,"type":"confidence_interval","upper":1256.6612255140012}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1246.7276612449236,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9d5796d5a175c67e0d98","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b0130e595d789a1491e426e5","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1245.589759515466,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1053,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=52","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1053,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1245.589759515466,"rating_lower":1238.5217824444176,"rating_upper":1252.6577365865146,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.0045126626769,"vote_count":12453},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1238.5217824444176,"type":"confidence_interval","upper":1252.6577365865146}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1245.589759515466,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bfefe79bb624ce627567","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-140104f8a72830dfc591793e","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1251.5525820633393,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1937,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=936","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1937,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1251.5525820633393,"rating_lower":1244.1639609777997,"rating_upper":1258.9412031488791,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":14.211195302053834,"vote_count":8699},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1244.1639609777997,"type":"confidence_interval","upper":1258.9412031488791}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1251.5525820633393,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cfc86ce6c24c51dd1767","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cbc54a52ee8724500d3c999f","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1260.2362448052645,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1760,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=759","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1760,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1260.2362448052645,"rating_lower":1246.196706043233,"rating_upper":1274.2757835672958,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.31088418512988,"vote_count":1870},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1246.196706043233,"type":"confidence_interval","upper":1274.2757835672958}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1260.2362448052645,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e923fade5014e8e355e7","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b503c3f4fccbbd4b9605179f","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1226.2761926535172,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1380,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=379","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1380,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1226.2761926535172,"rating_lower":1204.1195291360627,"rating_upper":1248.432856170972,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":127.79461166708458,"vote_count":737},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1204.1195291360627,"type":"confidence_interval","upper":1248.432856170972}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1226.2761926535172,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8bd4109089d6e3389a21","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-45c6dd76bd4556d70033cc10","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1250.807278997042,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87857b0f0260cb8ef7e3173fb1b9cd1f1939b066","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1466,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=465","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1466,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1250.807278997042,"rating_lower":1239.6494112301898,"rating_upper":1261.9651467638944,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.4090453428481,"vote_count":3232},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1239.6494112301898,"type":"confidence_interval","upper":1261.9651467638944}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1250.807278997042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b3783051ae0f95b869db","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a80e88deb9cb6bd5bee51605","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1196.1628693715452,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1695,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=694","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1695,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1196.1628693715452,"rating_lower":1149.4528187991675,"rating_upper":1242.8729199439226,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":567.9688176586574,"vote_count":149},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1149.4528187991675,"type":"confidence_interval","upper":1242.8729199439226}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1196.1628693715452,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-74dbd1f9135dff651c14","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-170d69c4ea012e1b5280978f","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1307.7550228241437,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1216,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=215","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1216,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1307.7550228241437,"rating_lower":1287.5402045733529,"rating_upper":1327.9698410749345,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":106.37596183802589,"vote_count":1250},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1287.5402045733529,"type":"confidence_interval","upper":1327.9698410749345}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1307.7550228241437,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f632bd323920508de217","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a2dc0a80b1511baa9158e358","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1221.721067038316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1857,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=856","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1857,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1221.721067038316,"rating_lower":1200.4435893374311,"rating_upper":1242.998544739201,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":117.85393998570774,"vote_count":812},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1200.4435893374311,"type":"confidence_interval","upper":1242.998544739201}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1221.721067038316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5e20cbb843f6dab8bfec","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cc8548cdd8c39cfcb6e8d490","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":55,"rawValue":1254.8602796340763,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1761,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=760","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1761,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1254.8602796340763,"rating_lower":1243.633302309509,"rating_upper":1266.0872569586434,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.81175869108211,"vote_count":3019},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1243.633302309509,"type":"confidence_interval","upper":1266.0872569586434}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1254.8602796340763,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-545c960d2a412a2f3c65","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4dfc61902b108a6bd2b2153a","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":55,"rawValue":1244.6977423948695,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1567,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=566","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1567,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1244.6977423948695,"rating_lower":1236.295242118562,"rating_upper":1253.1002426711768,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.378958148141134,"vote_count":9671},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1236.295242118562,"type":"confidence_interval","upper":1253.1002426711768}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1244.6977423948695,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2f4e8afc46333df51123","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e5c7ccb724f96f6cde17976c","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1241.322690881333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1056,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=55","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1056,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1241.322690881333,"rating_lower":1235.0446361574436,"rating_upper":1247.6007456052223,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.260157131927182,"vote_count":23081},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1235.0446361574436,"type":"confidence_interval","upper":1247.6007456052223}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1241.322690881333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1c493432537a1130ce63","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-937a112201404671eb0d89d8","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1248.9791828179539,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1940,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=939","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1940,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1248.9791828179539,"rating_lower":1242.826304285556,"rating_upper":1255.1320613503517,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.855087871956224,"vote_count":16445},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1242.826304285556,"type":"confidence_interval","upper":1255.1320613503517}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1248.9791828179539,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ad18aecfa8f402626110","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-298949554b8bcdefad45f193","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1250.3729374040126,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1467,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=466","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1467,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1250.3729374040126,"rating_lower":1241.6740470170453,"rating_upper":1259.07182779098,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.698426430300454,"vote_count":6089},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1241.6740470170453,"type":"confidence_interval","upper":1259.07182779098}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1250.3729374040126,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f6af87e8ac598086d0c4","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7884fecb70f97dc5c11016d4","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":60,"rawValue":1223.56785471743,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b1a6e388b2bb54527d5c9947bf32fbf37cddcfe","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1383,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=382","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1383,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1223.56785471743,"rating_lower":1206.8090903142102,"rating_upper":1240.3266191206496,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":73.11185605052883,"vote_count":1364},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1206.8090903142102,"type":"confidence_interval","upper":1240.3266191206496}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1223.56785471743,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-457c235f401e466a1f0c","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e8bff0432686733cdc8f2c08","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"entity_recognition","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1277.5273176078551,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1670,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=669","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1670,"metadata":{"category":"entity_recognition","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1277.5273176078551,"rating_lower":1237.579393545658,"rating_upper":1317.4752416700521,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":415.42463719309467,"vote_count":187},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1237.579393545658,"type":"confidence_interval","upper":1317.4752416700521}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1277.5273176078551,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7071c05b833cee8b36df","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-934cb1ca366dc1e66b43dbb7","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1270.788610518163,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1349,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=348","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1349,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1270.788610518163,"rating_lower":1255.2156683040657,"rating_upper":1286.3615527322604,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":63.1313624650004,"vote_count":1556},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1255.2156683040657,"type":"confidence_interval","upper":1286.3615527322604}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.788610518163,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5959f4c42e969a023bcc","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e1f26a98017be70d4efaf4bb","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1324.3954494186592,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1208,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=207","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1208,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1324.3954494186592,"rating_lower":1304.4598291485877,"rating_upper":1344.3310696887304,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":103.45782008947002,"vote_count":1264},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1304.4598291485877,"type":"confidence_interval","upper":1344.3310696887304}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1324.3954494186592,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-16437802daf14599c6c8","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a76605ed7c95725b81d21b62","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1272.2559184358659,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1543,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=542","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1543,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1272.2559184358659,"rating_lower":1264.1262384758788,"rating_upper":1280.3855983958529,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.204843091321248,"vote_count":10636},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1264.1262384758788,"type":"confidence_interval","upper":1280.3855983958529}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1272.2559184358659,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f0a50703d1a63b064b08","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3d5257ba0675b0d8b22d6439","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1288.1532616900909,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1737,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=736","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1737,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1288.1532616900909,"rating_lower":1277.3320645542349,"rating_upper":1298.974458825947,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":30.48277045747472,"vote_count":3204},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1277.3320645542349,"type":"confidence_interval","upper":1298.974458825947}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1288.1532616900909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9e85f29f831eaf6da1d3","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1a3f446f18ef7ccac3c4e0cb","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1270.3556461076255,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1835,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=834","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1835,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1270.3556461076255,"rating_lower":1249.9550964982873,"rating_upper":1290.7561957169637,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":108.33968130051841,"vote_count":864},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1249.9550964982873,"type":"confidence_interval","upper":1290.7561957169637}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1270.3556461076255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-23e4e629572c1ff2c783","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c40ab8047e5ac32ed704917d","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1273.1663913122293,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1920,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=919","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1920,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1273.1663913122293,"rating_lower":1267.276880421828,"rating_upper":1279.0559022026303,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.029470351548616,"vote_count":17806},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1267.276880421828,"type":"confidence_interval","upper":1279.0559022026303}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1273.1663913122293,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-54fd4f6758855e888a18","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f50dff7287c226e2bfc2ee0a","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1264.8198799115057,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1037,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=36","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1037,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1264.8198799115057,"rating_lower":1258.8127078806244,"rating_upper":1270.8270519423868,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.393857253969076,"vote_count":25396},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1258.8127078806244,"type":"confidence_interval","upper":1270.8270519423868}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1264.8198799115057,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-24a4b8f0e357719c73c2","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-65d9a6a6989a5823895ca2af","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1280.88441786539,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba87b8e2942ab2331aa2f424cd210a7f10f2b3bd","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1446,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=445","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1446,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1280.88441786539,"rating_lower":1272.615428524064,"rating_upper":1289.153407206716,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.799536040479044,"vote_count":6687},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1272.615428524064,"type":"confidence_interval","upper":1289.153407206716}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1280.88441786539,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ca85566efde512621ea4","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-184e572eb87705ed13f8c713","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1287.8850369104794,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1819,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=818","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1819,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1287.8850369104794,"rating_lower":1254.6784280666554,"rating_upper":1321.0916457543033,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":287.0469064946219,"vote_count":313},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1254.6784280666554,"type":"confidence_interval","upper":1321.0916457543033}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.8850369104794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0cdcea73abb088a358f2","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7bfe573395c60eda4ac4ac44","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1287.7995597979207,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1911,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=910","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1911,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1287.7995597979207,"rating_lower":1279.1738170480044,"rating_upper":1296.425302547837,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.368537178355105,"vote_count":6104},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1279.1738170480044,"type":"confidence_interval","upper":1296.425302547837}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.7995597979207,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-abd217ccb44150e1e1b7","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8327ad83008d58ad242de7a0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1280.597168003692,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1028,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=27","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1028,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1280.597168003692,"rating_lower":1272.3958001176923,"rating_upper":1288.7985358896917,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.509607245862515,"vote_count":8627},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1272.3958001176923,"type":"confidence_interval","upper":1288.7985358896917}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1280.597168003692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5ff558ff53bd32e3a304","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-68f547aed5bbb50b329abfd3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1268.346504123658,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1351,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=350","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1351,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1268.346504123658,"rating_lower":1241.8144608302814,"rating_upper":1294.8785474170345,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":183.25051866478324,"vote_count":535},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1241.8144608302814,"type":"confidence_interval","upper":1294.8785474170345}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1268.346504123658,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-13e57349d76a67a31b3d","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4afe8cc190b11c0a7069cab1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1274.6474959039533,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1540,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=539","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1540,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1274.6474959039533,"rating_lower":1262.8934725598203,"rating_upper":1286.4015192480863,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.96473923660622,"vote_count":3301},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.8934725598203,"type":"confidence_interval","upper":1286.4015192480863}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1274.6474959039533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d89b7b7f3bf3be7ce12b","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6afc0b7da67994ac44855e8a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1290.1653710018306,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1440,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=439","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1440,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1290.1653710018306,"rating_lower":1277.232723347363,"rating_upper":1303.098018656298,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.5390259694062,"vote_count":2339},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1277.232723347363,"type":"confidence_interval","upper":1303.098018656298}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1290.1653710018306,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-66dd29e082bfc700e9fe","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5c03537c0fc538cf13c162ce","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1323.080457534263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1210,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=209","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1210,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1323.080457534263,"rating_lower":1289.1666839721101,"rating_upper":1356.994231096416,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":299.4029328204961,"vote_count":353},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1289.1666839721101,"type":"confidence_interval","upper":1356.994231096416}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1323.080457534263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e87f06481371b8cd46f3","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ef3dbfa86f0daed725861200","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1287.0197766302144,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f3815b835a206f491e0b9dd163d6c750e38d7fe","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1739,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=738","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1739,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1287.0197766302144,"rating_lower":1267.685496827598,"rating_upper":1306.3540564328307,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":97.31052522861911,"vote_count":963},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1267.685496827598,"type":"confidence_interval","upper":1306.3540564328307}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1287.0197766302144,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47059fc96c576062aacb","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b0de2aff4268cc30a13474a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1280.7069192307151,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b8ec69fea951c4de3da84fe2cf5fe71d050e528","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1027,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=26","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1027,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1280.7069192307151,"rating_lower":1267.887112379453,"rating_upper":1293.5267260819774,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":42.782561358805566,"vote_count":2446},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1267.887112379453,"type":"confidence_interval","upper":1293.5267260819774}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1280.7069192307151,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e92f3b3fe50e6a83b541","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bb9e381d9defad90f510c003","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1295.6609945028163,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b8ec69fea951c4de3da84fe2cf5fe71d050e528","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1438,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=437","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1438,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1295.6609945028163,"rating_lower":1270.5148283876215,"rating_upper":1320.8071606180108,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":164.60665070430312,"vote_count":598},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1270.5148283876215,"type":"confidence_interval","upper":1320.8071606180108}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1295.6609945028163,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-513253089936d7237bf5","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ba342c828e283e4d52b295d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1283.2063334985266,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b8ec69fea951c4de3da84fe2cf5fe71d050e528","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1915,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=914","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1915,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1283.2063334985266,"rating_lower":1267.9425551926374,"rating_upper":1298.4701118044159,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":60.649596688695695,"vote_count":1625},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1267.9425551926374,"type":"confidence_interval","upper":1298.4701118044159}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1283.2063334985266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d4c70b32c98d30d4c11f","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d24fd33a2541795b9b9f30fc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1269.0314203901398,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b8ec69fea951c4de3da84fe2cf5fe71d050e528","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1546,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=545","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1546,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1269.0314203901398,"rating_lower":1247.0864114224237,"rating_upper":1290.976429357856,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":125.36472238068303,"vote_count":788},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1247.0864114224237,"type":"confidence_interval","upper":1290.976429357856}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1269.0314203901398,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6662b15709a56713a1e3","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5ba48893b8a57c017cc64d22","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1271.7680370261946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b8ec69fea951c4de3da84fe2cf5fe71d050e528","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1753,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=752","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1753,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1271.7680370261946,"rating_lower":1231.681028716384,"rating_upper":1311.8550453360053,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":418.3223900706703,"vote_count":223},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1231.681028716384,"type":"confidence_interval","upper":1311.8550453360053}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1271.7680370261946,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9490aa330ca56162abdb","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5e7cda915be0b7bb11eb65c2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1247.2306170456689,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b8ec69fea951c4de3da84fe2cf5fe71d050e528","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1371,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=370","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1371,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1247.2306170456689,"rating_lower":1203.0700770755116,"rating_upper":1291.3911570158261,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":507.65955890254196,"vote_count":193},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1203.0700770755116,"type":"confidence_interval","upper":1291.3911570158261}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1247.2306170456689,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0f82349ec1594de354a1","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e3c87c421155e1c88aaea7e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1314.938580023343,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1005,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=4","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1005,"metadata":{"category":"overall","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1314.938580023343,"rating_lower":1306.4900525969326,"rating_upper":1323.3871074497536,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.580861856515398,"vote_count":6743},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1306.4900525969326,"type":"confidence_interval","upper":1323.3871074497536}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1314.938580023343,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-df86e89fa16ebc20047e","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-12ece2d189d010ffafbb78f4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"homework","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1337.8449552750146,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1711,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=710","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1711,"metadata":{"category":"homework","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1337.8449552750146,"rating_lower":1313.4311034293357,"rating_upper":1362.2588071206933,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":155.15880548605034,"vote_count":596},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1313.4311034293357,"type":"confidence_interval","upper":1362.2588071206933}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1337.8449552750146,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-64150ab5e39050fc22aa","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4b345e036b443f000b23df07","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"diagram","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1331.8903493484236,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1415,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=414","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1415,"metadata":{"category":"diagram","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1331.8903493484236,"rating_lower":1317.1701757787932,"rating_upper":1346.6105229180537,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":56.40656845070334,"vote_count":1724},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1317.1701757787932,"type":"confidence_interval","upper":1346.6105229180537}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1331.8903493484236,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-36b80917a2c252615618","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3b0067c954cdcd03adfb2b3d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"ocr","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1324.0644823924113,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1890,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=889","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1890,"metadata":{"category":"ocr","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1324.0644823924113,"rating_lower":1314.9192466412883,"rating_upper":1333.2097181435345,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.771764542436824,"vote_count":4769},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1314.9192466412883,"type":"confidence_interval","upper":1333.2097181435345}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1324.0644823924113,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-38dbe2dfdc9b2a85f6f1","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b7a5a00699b4796532fe7135","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"english","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1310.458199904866,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1519,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=518","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1519,"metadata":{"category":"english","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1310.458199904866,"rating_lower":1296.8826237936416,"rating_upper":1324.0337760160903,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":47.97559347996459,"vote_count":2224},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1296.8826237936416,"type":"confidence_interval","upper":1324.0337760160903}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1310.458199904866,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7bd97b90cc8c8be26c81","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5af1e6d57e0ba8ee3ae1e6a6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"creative_writing_vision","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1314.9617396608874,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1331,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=330","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1331,"metadata":{"category":"creative_writing_vision","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1314.9617396608874,"rating_lower":1285.2685759478024,"rating_upper":1344.6549033739725,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":229.51800668600183,"vote_count":407},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1285.2685759478024,"type":"confidence_interval","upper":1344.6549033739725}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1314.9617396608874,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fb1560c4488b917523c9","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ba48eec9a9563ea984753b5d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"humor","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1294.1556781586576,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1815,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=814","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1815,"metadata":{"category":"humor","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1294.1556781586576,"rating_lower":1252.8889027935957,"rating_upper":1335.4224535237195,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":443.30730290706856,"vote_count":216},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1252.8889027935957,"type":"confidence_interval","upper":1335.4224535237195}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1294.1556781586576,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-vision","benchmarkName":"Arena Vision","benchmarkVersion":null,"benchmarkVersionHint":"arena-vision@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8e7eec364df51c5fe24d","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a1169fd86519f57742bf022d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"vision","category":"chinese","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"model"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1355.347785454357,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"447f1a2f0f1f911ad440fd210af72dfb340affa5","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1195,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=vision;split=latest;row_idx=194","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1195,"metadata":{"category":"chinese","config":"vision","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1355.347785454357,"rating_lower":1317.9566107657683,"rating_upper":1392.7389601429456,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":363.9502620881729,"vote_count":266},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1317.9566107657683,"type":"confidence_interval","upper":1392.7389601429456}},"sourceSubjectType":"model","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"rating","value":1355.347785454357,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6d55c247eaae010df686","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a0077754bce50747b48f8f16","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1666.4803631887767,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae739c6b187b8f120c33346c557fc533368c4bd1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2269,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=281","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2269,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1666.4803631887767,"rating_lower":1643.9876765377558,"rating_upper":1688.9730498397976,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":131.70021504736408,"vote_count":1001},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1643.9876765377558,"type":"confidence_interval","upper":1688.9730498397976}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1666.4803631887767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a98094be5e46f7d51b30","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-28d118fc4b7a613f5e3889aa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1623.1497118889843,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae739c6b187b8f120c33346c557fc533368c4bd1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2107,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=119","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2107,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1623.1497118889843,"rating_lower":1611.178227517959,"rating_upper":1635.1211962600096,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":37.30781578957733,"vote_count":3229},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1611.178227517959,"type":"confidence_interval","upper":1635.1211962600096}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1623.1497118889843,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6ff65cfcb8e214e8f05c","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-600c9682408cecc45a896182","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1626.2590219836861,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae739c6b187b8f120c33346c557fc533368c4bd1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2155,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=167","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2155,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1626.2590219836861,"rating_lower":1618.0557344653528,"rating_upper":1634.4623095020195,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.51780488858292,"vote_count":8382},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1618.0557344653528,"type":"confidence_interval","upper":1634.4623095020195}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1626.2590219836861,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-71c5c07669958a4f114f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-577eab6ceb0e07bd38e7c9e5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1629.661824704901,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae739c6b187b8f120c33346c557fc533368c4bd1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2390,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=402","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2390,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1629.661824704901,"rating_lower":1620.3047478120698,"rating_upper":1639.018901597732,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.79209333357715,"vote_count":6222},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1620.3047478120698,"type":"confidence_interval","upper":1639.018901597732}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1629.661824704901,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cab35f8847e7951965a8","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-14c8cea38b26a7e5cda2f51e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1626.2590219836861,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae739c6b187b8f120c33346c557fc533368c4bd1","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1993,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=5","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1993,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1626.2590219836861,"rating_lower":1618.0557344653528,"rating_upper":1634.4623095020195,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.51780488858292,"vote_count":8382},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1618.0557344653528,"type":"confidence_interval","upper":1634.4623095020195}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1626.2590219836861,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9b89f8fe1c5667d4d198","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ab11c844fd4721ce2eb18fdd","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":87,"rawValue":1324.1907410858214,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4556d87a8639e390f109a31fd1339bad5e31cf08","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2471,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=483","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2471,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1324.1907410858214,"rating_lower":1318.0433180335792,"rating_upper":1330.3381641380636,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.837619494879808,"vote_count":14482},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1318.0433180335792,"type":"confidence_interval","upper":1330.3381641380636}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1324.1907410858214,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-717d3b5f9ab120fdc76f","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-94d49cd628fc8feb80db7f9f","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":94,"rawValue":1327.4094602355049,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4556d87a8639e390f109a31fd1339bad5e31cf08","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2081,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=93","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2081,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1327.4094602355049,"rating_lower":1322.361828932383,"rating_upper":1332.4570915386264,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.632527631169323,"vote_count":26652},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1322.361828932383,"type":"confidence_interval","upper":1332.4570915386264}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1327.4094602355049,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3ecb54da29e748cd715b","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5f0ab35fa131fb97f8e329d2","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":94,"rawValue":1327.4094602355049,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4556d87a8639e390f109a31fd1339bad5e31cf08","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2243,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=255","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2243,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1327.4094602355049,"rating_lower":1322.361828932383,"rating_upper":1332.4570915386264,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":6.632527631169323,"vote_count":26652},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1322.361828932383,"type":"confidence_interval","upper":1332.4570915386264}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1327.4094602355049,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1e99ffd83ade5faddd65","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5263c8c20a0d0f68ba1cf340","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":95,"rawValue":1315.340717059536,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4556d87a8639e390f109a31fd1339bad5e31cf08","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2362,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=374","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2362,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1315.340717059536,"rating_lower":1306.025302409601,"rating_upper":1324.656131709471,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.589582278678392,"vote_count":10150},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1306.025302409601,"type":"confidence_interval","upper":1324.656131709471}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1315.340717059536,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-79619b1b84dac0883358","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-58ae30cd39bde23605404ab8","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1481.8384219267882,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2308,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=320","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2308,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1481.8384219267882,"rating_lower":1470.2168426534877,"rating_upper":1493.4600012000888,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.15880583647766,"vote_count":6492},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1470.2168426534877,"type":"confidence_interval","upper":1493.4600012000888}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1481.8384219267882,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fe0051019df8f5671341","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e4fa4a063a4cecad95f16040","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1468.0777023872317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2191,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=203","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2191,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1468.0777023872317,"rating_lower":1460.8400211740902,"rating_upper":1475.3153836003735,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.636493787430219,"vote_count":12925},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.8400211740902,"type":"confidence_interval","upper":1475.3153836003735}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1468.0777023872317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b8cad757efed04453a6c","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cc2b38dc109e18c6a0abe327","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1456.3995394700153,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2426,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=438","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2426,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1456.3995394700153,"rating_lower":1447.3979537377504,"rating_upper":1465.4011252022804,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.09317045358144,"vote_count":6405},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.3979537377504,"type":"confidence_interval","upper":1465.4011252022804}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1456.3995394700153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3a570e648a1b7a672fd4","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a4f22b81df54bf4ba66e0f79","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1468.0777023872317,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2029,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=41","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2029,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1468.0777023872317,"rating_lower":1460.8400211740902,"rating_upper":1475.3153836003735,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.636493787430219,"vote_count":12925},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.8400211740902,"type":"confidence_interval","upper":1475.3153836003735}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1468.0777023872317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-91aa37e3ce13bed6c95a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b851d201941cab101f62c206","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1511.5803414416328,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2301,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=313","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2301,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1511.5803414416328,"rating_lower":1499.0688441372863,"rating_upper":1524.0918387459792,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.749510043786586,"vote_count":6045},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1499.0688441372863,"type":"confidence_interval","upper":1524.0918387459792}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1511.5803414416328,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-08b3597f6ad7fca78553","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aa8acf84715874de7476ed22","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1494.3932987922449,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2022,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=34","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2022,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1494.3932987922449,"rating_lower":1486.0529726767943,"rating_upper":1502.7336249076957,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.107974849903307,"vote_count":10617},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1486.0529726767943,"type":"confidence_interval","upper":1502.7336249076957}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1494.3932987922449,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47f80651d5f3c7a39734","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1c7139c8285cb3036d1cef57","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1494.3932987922449,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2184,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=196","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2184,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1494.3932987922449,"rating_lower":1486.0529726767943,"rating_upper":1502.7336249076957,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.107974849903307,"vote_count":10617},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1486.0529726767943,"type":"confidence_interval","upper":1502.7336249076957}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1494.3932987922449,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5ee9038a2d9006bd19dc","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5cbe7ca4162f9066f0611be6","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1480.0731797318567,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2707037f7773a4c674d8bf72985f7eea3895aad9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2421,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=433","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2421,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1480.0731797318567,"rating_lower":1469.3332743696117,"rating_upper":1490.8130850941018,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":30.026501017948956,"vote_count":4562},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1469.3332743696117,"type":"confidence_interval","upper":1490.8130850941018}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1480.0731797318567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-af0e29cede319b09a509","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-587d9786185483ee03d1bbff","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1539.349142690242,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2118,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=130","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2118,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1539.349142690242,"rating_lower":1529.8648572716734,"rating_upper":1548.833428108811,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.41601826272419,"vote_count":7086},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1529.8648572716734,"type":"confidence_interval","upper":1548.833428108811}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1539.349142690242,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5da5b36b0abb9587770e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-90cead1b76a6cae8ec24c9f6","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1550.0478852675992,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2284,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=296","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2284,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1550.0478852675992,"rating_lower":1536.5801788816862,"rating_upper":1563.5155916535118,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":47.21620711383401,"vote_count":2267},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1536.5801788816862,"type":"confidence_interval","upper":1563.5155916535118}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1550.0478852675992,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2d74cad6b8c351df9d74","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-912e69143a113afa7e500601","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1536.273082073974,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2172,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=184","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2172,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1536.273082073974,"rating_lower":1530.6421346174996,"rating_upper":1541.904029530448,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.254043772840488,"vote_count":18226},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.6421346174996,"type":"confidence_interval","upper":1541.904029530448}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1536.273082073974,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-11d04f5d6d9fee86eae8","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-177a9e2d3bb1bcd10ebe2c0b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1536.273082073974,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2010,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=22","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2010,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1536.273082073974,"rating_lower":1530.6421346174996,"rating_upper":1541.904029530448,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.254043772840488,"vote_count":18226},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.6421346174996,"type":"confidence_interval","upper":1541.904029530448}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1536.273082073974,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f4924180c484e89c0d2f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-960bb971202ae5582b215be7","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1528.6853512923549,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2408,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=420","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2408,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1528.6853512923549,"rating_lower":1522.504281827652,"rating_upper":1534.8664207570578,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.945601791087059,"vote_count":14574},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1522.504281827652,"type":"confidence_interval","upper":1534.8664207570578}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1528.6853512923549,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-caeb42a72216c4c1fa2c","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-97f979395a32687b8b778e53","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1541.2452138974932,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2117,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=129","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2117,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1541.2452138974932,"rating_lower":1531.8858477094504,"rating_upper":1550.6045800855354,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.803247289800584,"vote_count":7006},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1531.8858477094504,"type":"confidence_interval","upper":1550.6045800855354}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1541.2452138974932,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c714205119f08d72290a","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fa181f8953d55b9c39815825","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1559.2067277690508,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2281,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=293","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2281,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1559.2067277690508,"rating_lower":1545.151107444413,"rating_upper":1573.262348093689,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.428499414365525,"vote_count":2169},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1545.151107444413,"type":"confidence_interval","upper":1573.262348093689}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1559.2067277690508,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-653e3f1ad07fe98619b3","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aa64bfe98c47dc4f37046f8d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1545.6803042848985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2167,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=179","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2167,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1545.6803042848985,"rating_lower":1539.903108395307,"rating_upper":1551.45750017449,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.688363953536369,"vote_count":17015},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1539.903108395307,"type":"confidence_interval","upper":1551.45750017449}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1545.6803042848985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b689cdce86af4628f7e5","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-52afa7a513e9071161e9a293","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1545.6803042848985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2005,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=17","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2005,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1545.6803042848985,"rating_lower":1539.903108395307,"rating_upper":1551.45750017449,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.688363953536369,"vote_count":17015},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1539.903108395307,"type":"confidence_interval","upper":1551.45750017449}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1545.6803042848985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a5d95e293758290170ce","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-24138e48c33a9148637f5175","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1539.2014452376384,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc003364119bfc786e8bc5ff618aa624b356dae5","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2404,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=416","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2404,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1539.2014452376384,"rating_lower":1532.830314925096,"rating_upper":1545.5725755501805,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.566637143349654,"vote_count":13422},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1532.830314925096,"type":"confidence_interval","upper":1545.5725755501805}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1539.2014452376384,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d3a3ecd396641bacd9aa","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5e64b757f820197935ea1c6f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1564.300417312133,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2114,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=126","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2114,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1564.300417312133,"rating_lower":1554.8209680856467,"rating_upper":1573.7798665386194,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.392143930700914,"vote_count":6430},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1554.8209680856467,"type":"confidence_interval","upper":1573.7798665386194}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1564.300417312133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4cbf4eed4c62f0d90b7d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-77985bc120238c8c7809d8f5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1559.4957732437688,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2280,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=292","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2280,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1559.4957732437688,"rating_lower":1543.857778713058,"rating_upper":1575.1337677744796,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":63.65989702275379,"vote_count":1707},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1543.857778713058,"type":"confidence_interval","upper":1575.1337677744796}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1559.4957732437688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-41f11ea575e787c4bc35","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b26239b3f17451063c7b5469","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1557.5168669531042,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2002,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=14","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2002,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1557.5168669531042,"rating_lower":1551.2878537015315,"rating_upper":1563.7458802046767,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.10048731467475,"vote_count":14527},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1551.2878537015315,"type":"confidence_interval","upper":1563.7458802046767}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1557.5168669531042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-08058f5c0e6acf14be0c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-60fffaf7a631b0fc50e1cf1c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1557.5168669531042,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2164,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=176","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2164,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1557.5168669531042,"rating_lower":1551.2878537015315,"rating_upper":1563.7458802046767,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.10048731467475,"vote_count":14527},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1551.2878537015315,"type":"confidence_interval","upper":1563.7458802046767}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1557.5168669531042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2b5cc3c78a69388ae846","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-150a44ac5b383123443eccd9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1554.6020279194017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2401,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=413","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2401,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1554.6020279194017,"rating_lower":1547.7529761629423,"rating_upper":1561.4510796758611,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.21137910159453,"vote_count":11452},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1547.7529761629423,"type":"confidence_interval","upper":1561.4510796758611}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1554.6020279194017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c4eb6ac955ff34170276","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5635c3b901200f66c832fb63","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1575.8252930757917,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2110,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=122","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2110,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1575.8252930757917,"rating_lower":1566.0945974335173,"rating_upper":1585.5559887180661,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.648562460828646,"vote_count":5997},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1566.0945974335173,"type":"confidence_interval","upper":1585.5559887180661}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1575.8252930757917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-74eeaddefca09182bc13","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-54a73eef52d403897d2fe5c0","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1555.0847118649995,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2282,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=294","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2282,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1555.0847118649995,"rating_lower":1540.1606483783821,"rating_upper":1570.0087753516168,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.97997098205061,"vote_count":1862},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1540.1606483783821,"type":"confidence_interval","upper":1570.0087753516168}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1555.0847118649995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c9e8ba1bcf23a3a17002","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d45661d9c661248c82c0fe16","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1557.7123580122975,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2400,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=412","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2400,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1557.7123580122975,"rating_lower":1550.8819913190805,"rating_upper":1564.5427247055143,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.14484167121101,"vote_count":11892},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1550.8819913190805,"type":"confidence_interval","upper":1564.5427247055143}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1557.7123580122975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e448b779f2bd3be5ce41","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-374a996952e57506cf7fa831","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1556.695153981578,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2165,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=177","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2165,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1556.695153981578,"rating_lower":1550.474680333418,"rating_upper":1562.9156276297379,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.07281197419143,"vote_count":15099},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1550.474680333418,"type":"confidence_interval","upper":1562.9156276297379}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1556.695153981578,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f1f3e9bcd14de2c02ced","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-023c2872d86f9fb16d6a934a","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1556.695153981578,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"440b77fa8596ad6d3936d0e0a86c0937059c29ee","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2003,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=15","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2003,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1556.695153981578,"rating_lower":1550.474680333418,"rating_upper":1562.9156276297379,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.07281197419143,"vote_count":15099},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1550.474680333418,"type":"confidence_interval","upper":1562.9156276297379}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1556.695153981578,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-508c65ef1d3657a95c0c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-eb7137e9b3ed0fb4e0e3f3cb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1541.3594837146388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2403,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=415","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2403,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1541.3594837146388,"rating_lower":1533.5514606336571,"rating_upper":1549.1675067956205,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.870331371175876,"vote_count":8057},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1533.5514606336571,"type":"confidence_interval","upper":1549.1675067956205}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1541.3594837146388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1891987dce6023f2d905","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7c8e3dff311f15d4352e9081","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1539.1006570198726,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2168,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=180","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2168,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1539.1006570198726,"rating_lower":1532.0983673569983,"rating_upper":1546.102946682747,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.763916733574824,"vote_count":10638},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1532.0983673569983,"type":"confidence_interval","upper":1546.102946682747}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1539.1006570198726,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b6dd10a25d5cae602846","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4902314580f9f5c84dc173d5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1539.1006570198726,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2006,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=18","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2006,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1539.1006570198726,"rating_lower":1532.0983673569983,"rating_upper":1546.102946682747,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.763916733574824,"vote_count":10638},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1532.0983673569983,"type":"confidence_interval","upper":1546.102946682747}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1539.1006570198726,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f8a0fe012894e4bcb01c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b4ee53a2cc752a9d17d8ede3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1527.232355179835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2295,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=307","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2295,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1527.232355179835,"rating_lower":1510.126231047322,"rating_upper":1544.3384793123482,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":76.17405170676147,"vote_count":1348},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1510.126231047322,"type":"confidence_interval","upper":1544.3384793123482}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1527.232355179835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1a15a1f2189306fbb63e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b2d7e993f8a3882270ffbaa5","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1562.9855618932615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2398,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=410","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2398,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1562.9855618932615,"rating_lower":1554.9271054025337,"rating_upper":1571.0440183839892,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.90470314639973,"vote_count":7595},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1554.9271054025337,"type":"confidence_interval","upper":1571.0440183839892}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1562.9855618932615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf6c9d6de9824ec598fc","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-95d113d47255e89b36ca5c6f","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1563.4925116674042,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2163,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=175","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2163,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1563.4925116674042,"rating_lower":1556.3726855065172,"rating_upper":1570.6123378282912,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.19600884126903,"vote_count":11699},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1556.3726855065172,"type":"confidence_interval","upper":1570.6123378282912}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1563.4925116674042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c5db9540cc03c920a6ac","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-58ca249b8e5d8a3c1df4118f","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":14,"rawValue":1563.4925116674042,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2001,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=13","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2001,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1563.4925116674042,"rating_lower":1556.3726855065172,"rating_upper":1570.6123378282912,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.19600884126903,"vote_count":11699},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1556.3726855065172,"type":"confidence_interval","upper":1570.6123378282912}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1563.4925116674042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3d2c70d37cccee100314","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3ebe46ae06ab0a034733b772","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1554.484277741286,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d9b65d228bbe8e09aab0933a5d4d0861881ab8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2283,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=295","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2283,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1554.484277741286,"rating_lower":1535.9444904986944,"rating_upper":1573.024064983878,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":89.4773905030327,"vote_count":1228},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1535.9444904986944,"type":"confidence_interval","upper":1573.024064983878}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1554.484277741286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f611780d5489a8e39d1d","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-40003d809af84f318424bf26","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1680.8083757859047,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2387,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=399","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2387,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1680.8083757859047,"rating_lower":1671.2132230954132,"rating_upper":1690.4035284763963,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.96666460613295,"vote_count":5893},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1671.2132230954132,"type":"confidence_interval","upper":1690.4035284763963}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1680.8083757859047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0947b55adff98c083553","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-893e236fc0d016763659de31","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1662.5934919996598,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1991,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=3","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1991,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1662.5934919996598,"rating_lower":1654.5162446083464,"rating_upper":1670.6707393909733,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.983632642114806,"vote_count":8659},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1654.5162446083464,"type":"confidence_interval","upper":1670.6707393909733}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1662.5934919996598,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d267307ae09ef6588ce7","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3d744cca97c9e64fe28a293e","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1648.1123955670514,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2271,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=283","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2271,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1648.1123955670514,"rating_lower":1626.2049447925106,"rating_upper":1670.0198463415927,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":124.93597402463479,"vote_count":999},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1626.2049447925106,"type":"confidence_interval","upper":1670.0198463415927}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1648.1123955670514,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-47457ebb5f424e3778cf","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3cff507069f9de4457484217","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1662.5934919996598,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2153,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=165","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2153,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1662.5934919996598,"rating_lower":1654.5162446083464,"rating_upper":1670.6707393909733,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.983632642114806,"vote_count":8659},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1654.5162446083464,"type":"confidence_interval","upper":1670.6707393909733}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1662.5934919996598,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c408ace9392b4b1c7778","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f24237a90d7cf973a7100983","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1685.0219335164438,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2268,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=280","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2268,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1685.0219335164438,"rating_lower":1659.791719783184,"rating_upper":1710.2521472497037,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":165.7088399846358,"vote_count":840},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1659.791719783184,"type":"confidence_interval","upper":1710.2521472497037}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1685.0219335164438,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2919b3db2dde615a183e","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ec0e93df7f6c17f2d0ea1747","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1709.1575588894566,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2385,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=397","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2385,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1709.1575588894566,"rating_lower":1698.632323598983,"rating_upper":1719.68279417993,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.83815318364102,"vote_count":5340},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1698.632323598983,"type":"confidence_interval","upper":1719.68279417993}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1709.1575588894566,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b6f1f305e79c1b7a09bb","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bef00c97fc9cff9a55725ddc","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1690.6395713997163,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2150,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=162","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2150,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1690.6395713997163,"rating_lower":1681.9304587549568,"rating_upper":1699.3486840444757,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.744749742079133,"vote_count":8116},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1681.9304587549568,"type":"confidence_interval","upper":1699.3486840444757}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1690.6395713997163,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-08fb0799d66fec30f831","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7b7365af989972b228c0ce78","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1690.6395713997163,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1988,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=0","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1988,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1690.6395713997163,"rating_lower":1681.9304587549568,"rating_upper":1699.3486840444757,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.744749742079133,"vote_count":8116},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1681.9304587549568,"type":"confidence_interval","upper":1699.3486840444757}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1690.6395713997163,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e675cf0ea0277b2a0be9","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-13243749e6268935a9ad567d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":1,"rawValue":1664.3882751396995,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facde5ecb28cee171a6045ea686668545ffa7595","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2106,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=118","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2106,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1664.3882751396995,"rating_lower":1648.6627342178326,"rating_upper":1680.1138160615665,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.37466827787755,"vote_count":1765},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1648.6627342178326,"type":"confidence_interval","upper":1680.1138160615665}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1664.3882751396995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-62f3c1f7303b539addcd","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6a682c7770fdc50911c18f94","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1383.8541311856036,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2452,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=464","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2452,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1383.8541311856036,"rating_lower":1373.957898399195,"rating_upper":1393.750363972012,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.49433117314787,"vote_count":4983},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1373.957898399195,"type":"confidence_interval","upper":1393.750363972012}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1383.8541311856036,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6b80bb9290623e1bc9c2","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-28f6f01740734984f3d1ec13","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":74,"rawValue":1385.5090474921917,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2061,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=73","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2061,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1385.5090474921917,"rating_lower":1378.2069202003247,"rating_upper":1392.8111747840587,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.880420297462564,"vote_count":15198},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1378.2069202003247,"type":"confidence_interval","upper":1392.8111747840587}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1385.5090474921917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1e3c9ffca8e6fb8ba5c3","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1213c9ffedef880964eb0166","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":74,"rawValue":1385.5090474921917,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2223,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=235","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2223,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1385.5090474921917,"rating_lower":1378.2069202003247,"rating_upper":1392.8111747840587,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.880420297462564,"vote_count":15198},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1378.2069202003247,"type":"confidence_interval","upper":1392.8111747840587}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1385.5090474921917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-07db2e2322f938633be0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-77146f8e9f1de134e2e4d82c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":75,"rawValue":1390.2739047408186,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2342,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=354","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2342,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1390.2739047408186,"rating_lower":1379.5607623801732,"rating_upper":1400.9870471014638,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.87704010288289,"vote_count":10192},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1379.5607623801732,"type":"confidence_interval","upper":1400.9870471014638}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1390.2739047408186,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-01b2489c8fea0321b6e3","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-69d185ca5a2ef26097308d2a","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":67,"rawValue":1385.519914415382,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2451,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=463","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2451,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1385.519914415382,"rating_lower":1374.4029148056486,"rating_upper":1396.6369140251152,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.17206954218466,"vote_count":3969},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1374.4029148056486,"type":"confidence_interval","upper":1396.6369140251152}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1385.519914415382,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-73803184342ed9bed6eb","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7bbf8d0e78009fd9df959e65","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1392.3609892159568,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2056,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=68","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2056,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1392.3609892159568,"rating_lower":1384.5976540712047,"rating_upper":1400.1243243607091,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.689188764713766,"vote_count":12802},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1384.5976540712047,"type":"confidence_interval","upper":1400.1243243607091}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1392.3609892159568,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3bb43539997495d25dd7","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-254d6ba36156ba935d137875","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1392.3609892159568,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2218,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=230","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2218,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1392.3609892159568,"rating_lower":1384.5976540712047,"rating_upper":1400.1243243607091,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.689188764713766,"vote_count":12802},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1384.5976540712047,"type":"confidence_interval","upper":1400.1243243607091}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1392.3609892159568,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-98254c4707da7aba5201","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f0426ac7a01edf476184248e","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929-high-32k","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":70,"rawValue":1398.7746311206224,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2bbaf51252538808ce6aa6f0d32473c8fa3c45","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2337,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=349","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2337,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1398.7746311206224,"rating_lower":1387.7940334107643,"rating_upper":1409.7552288304805,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.38743162264427,"vote_count":8826},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1387.7940334107643,"type":"confidence_interval","upper":1409.7552288304805}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1398.7746311206224,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-029ac6273ba99e2f2701","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4d5049a788b9ba2a9ed9ba6a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1529.1743836735386,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a5b99e04a9f66ec9f65254a762dc242dddc4026","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2122,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=134","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2122,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1529.1743836735386,"rating_lower":1519.852345185243,"rating_upper":1538.4964221618347,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.62171889208757,"vote_count":7334},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1519.852345185243,"type":"confidence_interval","upper":1538.4964221618347}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1529.1743836735386,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2c4c1c0ba108c413b585","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dffb3c16b8b5a3da859c56d0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1522.467769386207,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a5b99e04a9f66ec9f65254a762dc242dddc4026","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2174,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=186","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2174,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1522.467769386207,"rating_lower":1517.0516231817808,"rating_upper":1527.883915590633,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.636328040194705,"vote_count":20307},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1517.0516231817808,"type":"confidence_interval","upper":1527.883915590633}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1522.467769386207,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4415924f76e2a5a6a131","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8f7093e28e46cdd10aef5146","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1522.467769386207,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a5b99e04a9f66ec9f65254a762dc242dddc4026","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2012,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=24","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2012,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1522.467769386207,"rating_lower":1517.0516231817808,"rating_upper":1527.883915590633,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.636328040194705,"vote_count":20307},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1517.0516231817808,"type":"confidence_interval","upper":1527.883915590633}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1522.467769386207,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dfe5ddfdc4f37d8564fa","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dfdf760e35ef66c0ddd9ca6f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1529.0990045958115,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a5b99e04a9f66ec9f65254a762dc242dddc4026","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2294,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=306","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2294,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1529.0990045958115,"rating_lower":1516.4553388526833,"rating_upper":1541.7426703389401,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.61499338812857,"vote_count":2554},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1516.4553388526833,"type":"confidence_interval","upper":1541.7426703389401}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1529.0990045958115,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cbaffae95a29ebaa01fe","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5362d11832dde929b5540b04","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1511.8758873505844,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a5b99e04a9f66ec9f65254a762dc242dddc4026","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2412,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=424","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2412,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1511.8758873505844,"rating_lower":1505.9551504782971,"rating_upper":1517.7966242228717,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.125472052965506,"vote_count":16310},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1505.9551504782971,"type":"confidence_interval","upper":1517.7966242228717}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1511.8758873505844,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b8d94fa5f1d871b70ca1","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-792a7dc521282a838d4c3bfa","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1542.606007160266,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd757f8104f5d74dfc2bcb17efdde48ec6ffd7ee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2116,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=128","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2116,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1542.606007160266,"rating_lower":1532.5578763791,"rating_upper":1552.6541379414325,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":26.28296616158289,"vote_count":4358},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1532.5578763791,"type":"confidence_interval","upper":1552.6541379414325}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1542.606007160266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e877ed284ed5e01e4a27","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1fb85aa22dfaed4c16b523d4","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1542.5206202560817,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd757f8104f5d74dfc2bcb17efdde48ec6ffd7ee","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2402,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=414","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2402,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1542.5206202560817,"rating_lower":1533.306688346424,"rating_upper":1551.7345521657394,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.100078433345963,"vote_count":5270},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1533.306688346424,"type":"confidence_interval","upper":1551.7345521657394}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1542.5206202560817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fbdc4655578fdd310c1d","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-af197701a3096a0769a28011","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1538.5842623430603,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd757f8104f5d74dfc2bcb17efdde48ec6ffd7ee","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2009,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=21","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2009,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1538.5842623430603,"rating_lower":1530.5931385523327,"rating_upper":1546.5753861337882,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.623387733516743,"vote_count":7267},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.5931385523327,"type":"confidence_interval","upper":1546.5753861337882}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1538.5842623430603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dd4f5b0b4f9827d6c7a3","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-083d4e9db9d3550d543301d5","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1538.5842623430603,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd757f8104f5d74dfc2bcb17efdde48ec6ffd7ee","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2171,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=183","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2171,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1538.5842623430603,"rating_lower":1530.5931385523327,"rating_upper":1546.5753861337882,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.623387733516743,"vote_count":7267},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.5931385523327,"type":"confidence_interval","upper":1546.5753861337882}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1538.5842623430603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0e623e675ca4a1dd5000","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f17813b4e08c037a68cd512a","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1522.3186008214707,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd757f8104f5d74dfc2bcb17efdde48ec6ffd7ee","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2297,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=309","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2297,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"anthropic","rating":1522.3186008214707,"rating_lower":1500.1907343785788,"rating_upper":1544.4464672643621,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":127.46263754715311,"vote_count":821},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1500.1907343785788,"type":"confidence_interval","upper":1544.4464672643621}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1522.3186008214707,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d72afead981cf67bb2c6","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-90d7ef73ec5a9cc015608e57","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":85,"rawValue":1335.3422141277977,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2469,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=481","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2469,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1335.3422141277977,"rating_lower":1325.3591863818945,"rating_upper":1345.3252418737006,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.9434885618435,"vote_count":4888},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1325.3591863818945,"type":"confidence_interval","upper":1345.3252418737006}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1335.3422141277977,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-621eb3a1ebaed7827888","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0ba1d0f00081e32e4e293cb8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":95,"rawValue":1324.5403629748857,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2244,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=256","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2244,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1324.5403629748857,"rating_lower":1316.4896732037319,"rating_upper":1332.5910527460398,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.87213343071873,"vote_count":8988},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1316.4896732037319,"type":"confidence_interval","upper":1332.5910527460398}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1324.5403629748857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5b6ed50f826c1af7f7e8","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0add8cef7053fa1ed22d8263","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":95,"rawValue":1324.5403629748857,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2082,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=94","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2082,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1324.5403629748857,"rating_lower":1316.4896732037319,"rating_upper":1332.5910527460398,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.87213343071873,"vote_count":8988},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1316.4896732037319,"type":"confidence_interval","upper":1332.5910527460398}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1324.5403629748857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3b300f392b84ac981a64","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6558aec4066dc1f58fc042dc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":98,"rawValue":1303.936052735041,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2365,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=377","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2365,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1303.936052735041,"rating_lower":1291.5132092518957,"rating_upper":1316.3588962181861,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.17407120840664,"vote_count":4075},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1291.5132092518957,"type":"confidence_interval","upper":1316.3588962181861}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1303.936052735041,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-58b673b384da051727e2","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7188981c01735b13638c149f","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":101,"rawValue":1271.7662103406499,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"062a3574674b00e015e80e544d9874f3cff53222","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2088,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=100","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2088,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1271.7662103406499,"rating_lower":1258.1639228256702,"rating_upper":1285.3684978556296,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.16457348011841,"vote_count":3871},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1258.1639228256702,"type":"confidence_interval","upper":1285.3684978556296}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1271.7662103406499,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5bb9bee21b8dd4b340d6","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-415725d521532d5f54a957a3","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":101,"rawValue":1271.7662103406499,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"062a3574674b00e015e80e544d9874f3cff53222","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2250,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=262","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2250,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1271.7662103406499,"rating_lower":1258.1639228256702,"rating_upper":1285.3684978556296,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.16457348011841,"vote_count":3871},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1258.1639228256702,"type":"confidence_interval","upper":1285.3684978556296}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1271.7662103406499,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2d26d592e8ad985a0148","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-25b9afc3f05b8c07ff66546a","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2 Exp","modelRef":"deepseek-v3.2-exp","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":102,"rawValue":1281.774112303272,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"062a3574674b00e015e80e544d9874f3cff53222","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2369,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=381","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2369,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1281.774112303272,"rating_lower":1266.9940677626505,"rating_upper":1296.5541568438935,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":56.8663434437866,"vote_count":3871},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1266.9940677626505,"type":"confidence_interval","upper":1296.5541568438935}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1281.774112303272,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8b1289835fc54c639c46","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-580610a6a21a693ec3f6fc58","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":80,"rawValue":1349.8597570092388,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2464,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=476","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2464,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1349.8597570092388,"rating_lower":1338.043703459691,"rating_upper":1361.6758105587867,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":36.34533858169859,"vote_count":3503},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1338.043703459691,"type":"confidence_interval","upper":1361.6758105587867}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1349.8597570092388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0b687aa0dcbd0912f10c","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1a093866b8689bca5df17e47","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":81,"rawValue":1368.5876510360758,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2348,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=360","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2348,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1368.5876510360758,"rating_lower":1354.6233293426108,"rating_upper":1382.551972729541,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":50.76255908513068,"vote_count":3139},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1354.6233293426108,"type":"confidence_interval","upper":1382.551972729541}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1368.5876510360758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-196d2e7ef6efb0588f46","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-66318a921cf66c773a7d40f1","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":82,"rawValue":1360.3128404975844,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2231,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=243","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2231,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1360.3128404975844,"rating_lower":1351.1062721795968,"rating_upper":1369.5194088155722,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.064768659542427,"vote_count":6656},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1351.1062721795968,"type":"confidence_interval","upper":1369.5194088155722}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1360.3128404975844,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ef4aad6c192029f609c9","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-36dd4787028b44ad36877292","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V3.2","modelRef":"deepseek-v3.2-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":82,"rawValue":1360.3128404975844,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a90f60caf4caff65c50a6ad59f85855435d72b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2069,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=81","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2069,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1360.3128404975844,"rating_lower":1351.1062721795968,"rating_upper":1369.5194088155722,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.064768659542427,"vote_count":6656},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1351.1062721795968,"type":"confidence_interval","upper":1369.5194088155722}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1360.3128404975844,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c601af0fbd36d34e1381","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-18e6328029aa1800670cc662","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1583.5629500923421,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2396,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=408","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2396,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1583.5629500923421,"rating_lower":1570.3497928716827,"rating_upper":1596.7761073130016,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":45.4482351333181,"vote_count":2446},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1570.3497928716827,"type":"confidence_interval","upper":1596.7761073130016}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1583.5629500923421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c9508be041801d492aff","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5f662fcc22159f7058520800","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1579.0993765147234,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2162,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=174","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2162,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1579.0993765147234,"rating_lower":1568.1251416450928,"rating_upper":1590.073611384354,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.351066507607246,"vote_count":3537},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1568.1251416450928,"type":"confidence_interval","upper":1590.073611384354}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1579.0993765147234,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e687211cf2605f51b5a9","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2f319475fe37c7243da64399","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1579.0993765147234,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2000,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=12","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2000,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1579.0993765147234,"rating_lower":1568.1251416450928,"rating_upper":1590.073611384354,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.351066507607246,"vote_count":3537},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1568.1251416450928,"type":"confidence_interval","upper":1590.073611384354}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1579.0993765147234,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cfe073ea59209f11acf6","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-54a7265602a4ad92799e0dd0","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1542.27813546253,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2286,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=298","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2286,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1542.27813546253,"rating_lower":1510.9245189825892,"rating_upper":1573.6317519424701,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":255.90519442131034,"vote_count":404},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1510.9245189825892,"type":"confidence_interval","upper":1573.6317519424701}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1542.27813546253,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b388fd01fb5b0f935f01","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e791bed498375b34c913012c","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1422.5202197873225,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2441,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=453","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2441,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1422.5202197873225,"rating_lower":1414.665856193438,"rating_upper":1430.374583381207,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.059270799052914,"vote_count":7622},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1414.665856193438,"type":"confidence_interval","upper":1430.374583381207}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1422.5202197873225,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b091b994655d043e9c90","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cafef44bd84b3006c259219b","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1431.069014492644,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2045,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=57","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2045,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1431.069014492644,"rating_lower":1423.9124164385569,"rating_upper":1438.2256125467313,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.332668160298319,"vote_count":9497},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1423.9124164385569,"type":"confidence_interval","upper":1438.2256125467313}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1431.069014492644,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-eacc4ecbb7b6d210df52","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9f187821ef3a99a08d5e1bc3","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1431.069014492644,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2207,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=219","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2207,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1431.069014492644,"rating_lower":1423.9124164385569,"rating_upper":1438.2256125467313,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.332668160298319,"vote_count":9497},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1423.9124164385569,"type":"confidence_interval","upper":1438.2256125467313}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1431.069014492644,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5adabb82bf80a40d2b9f","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-09696af3a40b9e12fd48aa0b","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":61,"rawValue":1422.9704052212535,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376ca9e2712aba24ca0ba14c6409a03399519c17","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2328,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=340","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2328,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1422.9704052212535,"rating_lower":1404.1785071461177,"rating_upper":1441.7623032963893,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":91.92742907041728,"vote_count":1078},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1404.1785071461177,"type":"confidence_interval","upper":1441.7623032963893}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1422.9704052212535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9b2519f1e53c610b411a","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d860309764ed1e66975eed7d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1439.2137888468844,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2431,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=443","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2431,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1439.2137888468844,"rating_lower":1431.9276190628373,"rating_upper":1446.4999586309318,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.819820177682033,"vote_count":9071},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1431.9276190628373,"type":"confidence_interval","upper":1446.4999586309318}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1439.2137888468844,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4f7124e2ee4971d8fff8","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6e2415f8885566438278d486","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1445.4554435536177,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2198,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=210","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2198,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1445.4554435536177,"rating_lower":1438.9308823313079,"rating_upper":1451.9800047759275,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.081701283466147,"vote_count":13178},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1438.9308823313079,"type":"confidence_interval","upper":1451.9800047759275}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1445.4554435536177,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1edd60586c6e16cae318","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5e148ba8679b650bfe451bd7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1445.4554435536177,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2036,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=48","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2036,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1445.4554435536177,"rating_lower":1438.9308823313079,"rating_upper":1451.9800047759275,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.081701283466147,"vote_count":13178},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1438.9308823313079,"type":"confidence_interval","upper":1451.9800047759275}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1445.4554435536177,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b0f7ee142443b129ad11","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4d791da771c410de0043dfdb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1436.7593568233792,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2321,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=333","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2321,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1436.7593568233792,"rating_lower":1420.2342290499437,"rating_upper":1453.284484596815,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":71.08753748895654,"vote_count":1411},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1420.2342290499437,"type":"confidence_interval","upper":1453.284484596815}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1436.7593568233792,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3b11ad92f92932534de8","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-901c6d7a4864baa30be23b90","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1588.6753688261556,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9120fe5af37aa41a61217ad7519ebb60845a84a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2395,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=407","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2395,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1588.6753688261556,"rating_lower":1573.6352242871785,"rating_upper":1603.7155133651324,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.88542824791858,"vote_count":1867},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1573.6352242871785,"type":"confidence_interval","upper":1603.7155133651324}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1588.6753688261556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-aee0be649a96aaf534fd","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d132cf5488672826af4c680b","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1563.003289949429,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9120fe5af37aa41a61217ad7519ebb60845a84a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2279,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=291","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2279,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1563.003289949429,"rating_lower":1526.3928548065778,"rating_upper":1599.6137250922802,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":348.9101468766313,"vote_count":314},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1526.3928548065778,"type":"confidence_interval","upper":1599.6137250922802}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1563.003289949429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ad75414cbb316787a403","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9756fa882cdf906b4fc4fb83","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1581.6418400239618,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9120fe5af37aa41a61217ad7519ebb60845a84a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2161,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=173","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2161,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1581.6418400239618,"rating_lower":1569.5312670080477,"rating_upper":1593.752413039876,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.17976076788653,"vote_count":2869},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1569.5312670080477,"type":"confidence_interval","upper":1593.752413039876}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1581.6418400239618,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f551b85c984d0176ab67","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2bf3c4625ef2ad25b6d0125c","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-high-20260813","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":12,"rawValue":1581.6418400239618,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9120fe5af37aa41a61217ad7519ebb60845a84a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1999,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=11","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1999,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1581.6418400239618,"rating_lower":1569.5312670080477,"rating_upper":1593.752413039876,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.17976076788653,"vote_count":2869},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1569.5312670080477,"type":"confidence_interval","upper":1593.752413039876}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1581.6418400239618,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-373ef80737e4a38e980d","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f725d92533b7fc2b04ecaf8d","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1463.8042124059687,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2192,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=204","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2192,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1463.8042124059687,"rating_lower":1457.2943201849207,"rating_upper":1470.3141046270166,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.031927897122168,"vote_count":12409},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.2943201849207,"type":"confidence_interval","upper":1470.3141046270166}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1463.8042124059687,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-eec1538b92be5ad0d0df","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-759c221f7ccbe17a9f116a56","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1463.8042124059687,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2030,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=42","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2030,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1463.8042124059687,"rating_lower":1457.2943201849207,"rating_upper":1470.3141046270166,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.031927897122168,"vote_count":12409},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.2943201849207,"type":"confidence_interval","upper":1470.3141046270166}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1463.8042124059687,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e9ee9830ac3fe3999525","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4069a7243578690f69f6fcce","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1454.6125978774453,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2427,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=439","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2427,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1454.6125978774453,"rating_lower":1447.4920778767685,"rating_upper":1461.733117878122,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.198580915902316,"vote_count":9804},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.4920778767685,"type":"confidence_interval","upper":1461.733117878122}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1454.6125978774453,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6aeadbbabdafb8974afb","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4dcd94fc55c902d7eef8ba26","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro-high-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1471.740281364178,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"700619ef73d194cd28fc310efc583c9d24faced6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2311,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=323","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2311,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"deepseek","rating":1471.740281364178,"rating_lower":1455.2771110021863,"rating_upper":1488.20345172617,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":70.55548191949389,"vote_count":1453},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1455.2771110021863,"type":"confidence_interval","upper":1488.20345172617}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1471.740281364178,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e98e7a7c8470d8179c28","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-68258c164a897ee3a74e6ddb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1283.6824608232687,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"861fc1ca72f628c9a839ff0ff38a5cd7547b0b11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2149,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=161","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2149,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1283.6824608232687,"rating_lower":1262.1365038794975,"rating_upper":1305.22841776704,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":120.84686633161893,"vote_count":984},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1262.1365038794975,"type":"confidence_interval","upper":1305.22841776704}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1283.6824608232687,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4775f896e714000286a3","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9b5a7fe96d3a2e447c3b9112","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":110,"rawValue":1234.8747938414745,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"861fc1ca72f628c9a839ff0ff38a5cd7547b0b11","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2377,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=389","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2377,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1234.8747938414745,"rating_lower":1217.8611577577944,"rating_upper":1251.8884299251545,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":75.35257471160348,"vote_count":2675},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1217.8611577577944,"type":"confidence_interval","upper":1251.8884299251545}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1234.8747938414745,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d4f6efb1c7965d43bcbe","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ba3f2ab2a1a251a9d8a46c28","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":111,"rawValue":1225.6866415067607,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"861fc1ca72f628c9a839ff0ff38a5cd7547b0b11","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2260,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=272","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2260,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1225.6866415067607,"rating_lower":1209.6959201290774,"rating_upper":1241.6773628844437,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":66.56407945887858,"vote_count":2675},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1209.6959201290774,"type":"confidence_interval","upper":1241.6773628844437}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1225.6866415067607,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d8f8e505553e56c5798d","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6d7d476ee582ccb3340e13fc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":111,"rawValue":1225.6866415067607,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"861fc1ca72f628c9a839ff0ff38a5cd7547b0b11","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2098,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=110","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2098,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1225.6866415067607,"rating_lower":1209.6959201290774,"rating_upper":1241.6773628844437,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":66.56407945887858,"vote_count":2675},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1209.6959201290774,"type":"confidence_interval","upper":1241.6773628844437}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1225.6866415067607,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0c640fdd5aeb0dc17425","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f9b38676175a0b0196e2e56b","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1480.5198401926805,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"244b66964f5b0235aabb348a6f9332e4262db7aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2131,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=143","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2131,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1480.5198401926805,"rating_lower":1471.5127923916853,"rating_upper":1489.5268879936757,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.11877645346088,"vote_count":7430},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1471.5127923916853,"type":"confidence_interval","upper":1489.5268879936757}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1480.5198401926805,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dcfbb5b26e722cf9b02c","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0fb57f64a5a3e9e8a14ea929","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1493.2614015377562,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"244b66964f5b0235aabb348a6f9332e4262db7aa","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2305,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=317","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2305,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1493.2614015377562,"rating_lower":1480.2076749301827,"rating_upper":1506.3151281453297,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.35809058457576,"vote_count":2310},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1480.2076749301827,"type":"confidence_interval","upper":1506.3151281453297}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1493.2614015377562,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1c429c518019bb1eda95","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6a70888551f7db985ef3e781","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1436.495850185246,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"244b66964f5b0235aabb348a6f9332e4262db7aa","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2432,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=444","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2432,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1436.495850185246,"rating_lower":1430.6883279405088,"rating_upper":1442.3033724299828,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":8.779819385652406,"vote_count":16453},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.6883279405088,"type":"confidence_interval","upper":1442.3033724299828}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1436.495850185246,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9542cc8750e6334f76dc","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6292b59e643df4352d8298f4","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1446.4556668949795,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"244b66964f5b0235aabb348a6f9332e4262db7aa","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2035,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=47","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2035,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1446.4556668949795,"rating_lower":1441.163813532886,"rating_upper":1451.747520257073,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.289863906660375,"vote_count":21573},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1441.163813532886,"type":"confidence_interval","upper":1451.747520257073}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1446.4556668949795,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f28aecea653717604cd8","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1ac05ddb51726e56cf7537eb","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1446.4556668949795,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"244b66964f5b0235aabb348a6f9332e4262db7aa","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2197,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=209","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2197,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1446.4556668949795,"rating_lower":1441.163813532886,"rating_upper":1451.747520257073,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.289863906660375,"vote_count":21573},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1441.163813532886,"type":"confidence_interval","upper":1451.747520257073}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1446.4556668949795,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-abe04229465d763dd386","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4b25372ff623d034fc5924bd","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1493.6890751917033,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2130,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=142","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2130,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1493.6890751917033,"rating_lower":1478.5551026346798,"rating_upper":1508.8230477487268,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":59.62243409272238,"vote_count":1656},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1478.5551026346798,"type":"confidence_interval","upper":1508.8230477487268}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1493.6890751917033,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-594738746637ac21a766","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d1df4677fb6a7ba6494d7505","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1519.0834325869387,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2300,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=312","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2300,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1519.0834325869387,"rating_lower":1497.8455006292115,"rating_upper":1540.3213645446663,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":117.41626681281404,"vote_count":924},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1497.8455006292115,"type":"confidence_interval","upper":1540.3213645446663}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1519.0834325869387,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9c27ee683e2efa49a821","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d734714bba1d517c289228ed","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1498.5856161377683,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2183,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=195","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2183,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1498.5856161377683,"rating_lower":1490.5949270202302,"rating_upper":1506.5763052553064,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.621579340945033,"vote_count":7559},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1490.5949270202302,"type":"confidence_interval","upper":1506.5763052553064}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1498.5856161377683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-545ca4b6f843619d94af","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b03bad554b878e035d88ac90","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1498.5856161377683,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2021,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=33","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2021,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1498.5856161377683,"rating_lower":1490.5949270202302,"rating_upper":1506.5763052553064,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":16.621579340945033,"vote_count":7559},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1490.5949270202302,"type":"confidence_interval","upper":1506.5763052553064}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1498.5856161377683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-74bb4fcce1812bbda160","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6d4148fa89a6874a3298cfe2","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1491.539311545379,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2418,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=430","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2418,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1491.539311545379,"rating_lower":1482.3876619832452,"rating_upper":1500.6909611075125,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.802313552580564,"vote_count":5509},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1482.3876619832452,"type":"confidence_interval","upper":1500.6909611075125}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1491.539311545379,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-690b41ba2623df9fdd8b","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2699ada5a6f6059a04a8f1b2","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1443.7875790280625,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad521bc0b4d90d0d112a96ecdad5414848a70287","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2137,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=149","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2137,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1443.7875790280625,"rating_lower":1426.5657444046096,"rating_upper":1461.0094136515154,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":77.2080612187766,"vote_count":1269},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1426.5657444046096,"type":"confidence_interval","upper":1461.0094136515154}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1443.7875790280625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ae5c3aca3e597f8dd2b9","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9c319f416b1ceb61c2b75edc","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1449.1411432227515,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad521bc0b4d90d0d112a96ecdad5414848a70287","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2034,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=46","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2034,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1449.1411432227515,"rating_lower":1406.58595276385,"rating_upper":1491.696333681653,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":471.4209677942503,"vote_count":228},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1406.58595276385,"type":"confidence_interval","upper":1491.696333681653}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1449.1411432227515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bc4b38956a6edfe2d029","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7e50452a9446a6e3e62f3e9f","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":47,"rawValue":1449.1411432227515,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad521bc0b4d90d0d112a96ecdad5414848a70287","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2196,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=208","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2196,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1449.1411432227515,"rating_lower":1406.58595276385,"rating_upper":1491.696333681653,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":471.4209677942503,"vote_count":228},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1406.58595276385,"type":"confidence_interval","upper":1491.696333681653}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1449.1411432227515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0510c4d86a7473c6405b","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-58d0de43e1f396e8647eca55","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Arena base presentations of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Arena base presentations of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":55,"rawValue":1425.1457138983424,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad521bc0b4d90d0d112a96ecdad5414848a70287","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2439,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=451","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2439,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1425.1457138983424,"rating_lower":1370.917916674537,"rating_upper":1479.3735111221476,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":765.5044942574133,"vote_count":140},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1370.917916674537,"type":"confidence_interval","upper":1479.3735111221476}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1425.1457138983424,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-776e2f5b7ac151925fa8","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c7c84fc177960e7bce2178cb","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1499.011477726488,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2128,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=140","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2128,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1499.011477726488,"rating_lower":1486.4057885004336,"rating_upper":1511.6171669525424,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":41.365379216834064,"vote_count":2415},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1486.4057885004336,"type":"confidence_interval","upper":1511.6171669525424}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1499.011477726488,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7e42d020b6e35e1185f5","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-47d22e7ff3cc325a44129991","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1523.8905440947433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2296,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=308","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2296,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1523.8905440947433,"rating_lower":1504.4672937853315,"rating_upper":1543.3137944041555,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":98.20817303825926,"vote_count":1061},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1504.4672937853315,"type":"confidence_interval","upper":1543.3137944041555}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1523.8905440947433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-62eb3d1692ca6f2e321b","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dc8479bbdaab4dd23c3566d7","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1489.741632583298,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2185,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=197","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2185,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1489.741632583298,"rating_lower":1482.5146630323366,"rating_upper":1496.9686021342595,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.596160034090026,"vote_count":10038},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1482.5146630323366,"type":"confidence_interval","upper":1496.9686021342595}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1489.741632583298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9e33df22bd9f87cfffac","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5dff0d1ff95ac289fa7d9aed","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1489.741632583298,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2023,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=35","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2023,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1489.741632583298,"rating_lower":1482.5146630323366,"rating_upper":1496.9686021342595,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.596160034090026,"vote_count":10038},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1482.5146630323366,"type":"confidence_interval","upper":1496.9686021342595}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1489.741632583298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c1bacc0705d3a51dae01","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-582a942f3438bb169ccc52eb","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high and medium presentations of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high and medium presentations of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1482.952749144996,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833ca4bb04de81a5f5fb8ce7e43572ee98d6c8ae","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2420,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=432","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2420,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1482.952749144996,"rating_lower":1474.7822853749446,"rating_upper":1491.1232129150474,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.377897651304224,"vote_count":7143},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1474.7822853749446,"type":"confidence_interval","upper":1491.1232129150474}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1482.952749144996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a084906c3d97e5d5c57e","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-91438fbdddd3fc145182ac2a","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1544.437428907048,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19d2fa5f0e3f9261c807cd13e178d1483b0bb056","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2115,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=127","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2115,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1544.437428907048,"rating_lower":1528.3263306688473,"rating_upper":1560.5485271452485,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":67.57002965713123,"vote_count":1412},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1528.3263306688473,"type":"confidence_interval","upper":1560.5485271452485}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1544.437428907048,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cb46e4e54722e35fd81f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ea6fb39cc8d1edfec6108c43","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1538.9071087861485,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19d2fa5f0e3f9261c807cd13e178d1483b0bb056","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2008,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=20","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2008,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1538.9071087861485,"rating_lower":1530.3617726593923,"rating_upper":1547.4524449129046,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.009124639282117,"vote_count":6464},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.3617726593923,"type":"confidence_interval","upper":1547.4524449129046}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1538.9071087861485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8545f11e087d5393946f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-af422b3a2ee242cd5598ec34","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":21,"rawValue":1538.9071087861485,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19d2fa5f0e3f9261c807cd13e178d1483b0bb056","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2170,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=182","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2170,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1538.9071087861485,"rating_lower":1530.3617726593923,"rating_upper":1547.4524449129046,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.009124639282117,"vote_count":6464},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1530.3617726593923,"type":"confidence_interval","upper":1547.4524449129046}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1538.9071087861485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-83595f96e63ab3c43fb6","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-42328637bee9bcf2afb9b6d8","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1529.8411576375765,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19d2fa5f0e3f9261c807cd13e178d1483b0bb056","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2406,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=418","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2406,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1529.8411576375765,"rating_lower":1519.5549354339396,"rating_upper":1540.1273798412135,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":27.543277739334336,"vote_count":4260},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1519.5549354339396,"type":"confidence_interval","upper":1540.1273798412135}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1529.8411576375765,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-629b7b49d680d6fb43a8","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-15f16344d960f590f18348c4","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":25,"rawValue":1535.4738557700312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19d2fa5f0e3f9261c807cd13e178d1483b0bb056","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2292,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=304","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2292,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1535.4738557700312,"rating_lower":1511.7473042920165,"rating_upper":1559.2004072480458,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":146.54569300762685,"vote_count":717},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1511.7473042920165,"type":"confidence_interval","upper":1559.2004072480458}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1535.4738557700312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0dfa83f21e5efa05a74b","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cfd8cd9df2232ee76509b63d","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1622.34702959783,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26ecae5b67dabb85478af0a1cc525b05de250047","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2274,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=286","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2274,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1622.34702959783,"rating_lower":1584.3613938649069,"rating_upper":1660.3326653307531,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":375.6147311175936,"vote_count":320},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1584.3613938649069,"type":"confidence_interval","upper":1660.3326653307531}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1622.34702959783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c5ef1ed41fc817bcf790","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8d41218517459103f97d11fe","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1587.044813853989,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26ecae5b67dabb85478af0a1cc525b05de250047","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1997,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=9","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1997,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1587.044813853989,"rating_lower":1573.9055863185263,"rating_upper":1600.1840413894517,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.941077930770916,"vote_count":2552},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1573.9055863185263,"type":"confidence_interval","upper":1600.1840413894517}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1587.044813853989,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5efd66941c88afb86da4","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6cc35a212addd733f84799d2","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1587.044813853989,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26ecae5b67dabb85478af0a1cc525b05de250047","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2159,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=171","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2159,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1587.044813853989,"rating_lower":1573.9055863185263,"rating_upper":1600.1840413894517,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.941077930770916,"vote_count":2552},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1573.9055863185263,"type":"confidence_interval","upper":1600.1840413894517}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1587.044813853989,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f231973daed6783bd5f4","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a1b0af00e339346a8a409c76","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Arena high presentations of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Arena high presentations of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":13,"rawValue":1581.4200321442174,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26ecae5b67dabb85478af0a1cc525b05de250047","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2397,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=409","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2397,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"google","rating":1581.4200321442174,"rating_lower":1565.6767148430781,"rating_upper":1597.163349445357,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.52029065342722,"vote_count":1785},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1565.6767148430781,"type":"confidence_interval","upper":1597.163349445357}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1581.4200321442174,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c09901dbed0cf1315714","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-88f7329fc657ddc33d907db3","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":51,"rawValue":1446.2888781758202,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9c6aa7383e12335122544cc7ca9265c417cf3f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2318,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=330","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2318,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1446.2888781758202,"rating_lower":1432.648038656873,"rating_upper":1459.929717694767,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.437979285183836,"vote_count":3706},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1432.648038656873,"type":"confidence_interval","upper":1459.929717694767}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1446.2888781758202,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d0362af8321f27328d34","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-09f1b20408c5c255a43663cf","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":52,"rawValue":1428.757810074879,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9c6aa7383e12335122544cc7ca9265c417cf3f1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2436,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=448","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2436,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1428.757810074879,"rating_lower":1358.667001657249,"rating_upper":1498.848618492509,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":1278.8686938857277,"vote_count":100},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1358.667001657249,"type":"confidence_interval","upper":1498.848618492509}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1428.757810074879,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dd31621a5421d848e0f4","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5b3d88c93ccad8293428b7bf","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1434.289758171793,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9c6aa7383e12335122544cc7ca9265c417cf3f1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2043,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=55","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2043,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1434.289758171793,"rating_lower":1421.9211144057476,"rating_upper":1446.6584019378386,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.82428440653998,"vote_count":3814},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1421.9211144057476,"type":"confidence_interval","upper":1446.6584019378386}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1434.289758171793,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f60e2e4c77dd4805ef6f","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-517b9633f55c60b599ef8722","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1434.289758171793,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9c6aa7383e12335122544cc7ca9265c417cf3f1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2205,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=217","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2205,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1434.289758171793,"rating_lower":1421.9211144057476,"rating_upper":1446.6584019378386,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":39.82428440653998,"vote_count":3814},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1421.9211144057476,"type":"confidence_interval","upper":1446.6584019378386}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1434.289758171793,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-27d743cb5133118e4548","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0158143df2becd022771cb97","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1454.1117867061448,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8a8838baec7f0e85db912da7ccac601893ae091","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2317,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=329","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2317,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1454.1117867061448,"rating_lower":1432.6072459296938,"rating_upper":1475.6163274825958,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":120.38272322869382,"vote_count":837},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1432.6072459296938,"type":"confidence_interval","upper":1475.6163274825958}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1454.1117867061448,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fb30c034ce237a6c9d86","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9f4495f7ceb70c192e4ee0c5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":55,"rawValue":1435.4584267142257,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8a8838baec7f0e85db912da7ccac601893ae091","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2204,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=216","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2204,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1435.4584267142257,"rating_lower":1427.1289228731778,"rating_upper":1443.7879305552733,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.06101209891198,"vote_count":7068},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1427.1289228731778,"type":"confidence_interval","upper":1443.7879305552733}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1435.4584267142257,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-48f012dbf11523a189f1","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-01d1a0b6426a9318afd6e4de","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":55,"rawValue":1435.4584267142257,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8a8838baec7f0e85db912da7ccac601893ae091","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2042,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=54","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2042,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1435.4584267142257,"rating_lower":1427.1289228731778,"rating_upper":1443.7879305552733,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.06101209891198,"vote_count":7068},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1427.1289228731778,"type":"confidence_interval","upper":1443.7879305552733}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1435.4584267142257,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-29c97d7baac635f01f92","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4edf74e45d86fecd18714dac","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1424.6289525106836,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8a8838baec7f0e85db912da7ccac601893ae091","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2440,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=452","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2440,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1424.6289525106836,"rating_lower":1415.6772161931335,"rating_upper":1433.5806888282336,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.86019578480481,"vote_count":6156},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1415.6772161931335,"type":"confidence_interval","upper":1433.5806888282336}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1424.6289525106836,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-04de47a077f0873890cd","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a82c78acead6d334c6fdba92","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1508.9259030653723,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b02c19f04a25fe4d63167ce43426399eb24fa76","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2018,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=30","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2018,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1508.9259030653723,"rating_lower":1501.6846407121186,"rating_upper":1516.167165418626,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.649991556898474,"vote_count":9296},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.6846407121186,"type":"confidence_interval","upper":1516.167165418626}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1508.9259030653723,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7e43bb39411808431891","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-634eb959a04e8bb89b7e1828","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1504.6689943291394,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b02c19f04a25fe4d63167ce43426399eb24fa76","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2415,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=427","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2415,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1504.6689943291394,"rating_lower":1496.5808175258887,"rating_upper":1512.7571711323899,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.029625216396525,"vote_count":7210},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1496.5808175258887,"type":"confidence_interval","upper":1512.7571711323899}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1504.6689943291394,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-18ebdfd624fc9d6c6d97","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-585f5b708fba1e957522bffb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1508.9259030653723,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b02c19f04a25fe4d63167ce43426399eb24fa76","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2180,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=192","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2180,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1508.9259030653723,"rating_lower":1501.6846407121186,"rating_upper":1516.167165418626,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.649991556898474,"vote_count":9296},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.6846407121186,"type":"confidence_interval","upper":1516.167165418626}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1508.9259030653723,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0f829d60412affe062dd","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bbffb41a3bda9f9591dcb679","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1519.7441056250213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b02c19f04a25fe4d63167ce43426399eb24fa76","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2299,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=311","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2299,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1519.7441056250213,"rating_lower":1499.875698120903,"rating_upper":1539.61251312914,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":102.76138185398808,"vote_count":1027},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1499.875698120903,"type":"confidence_interval","upper":1539.61251312914}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1519.7441056250213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6ba738806d19c5d2cfff","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c61e5792588c2e4ce409957b","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1593.532925524643,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f8c6893005929dad7dcce46742401ccb807215a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2394,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=406","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2394,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1593.532925524643,"rating_lower":1584.5115853041812,"rating_upper":1602.5542657451047,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.18585234726405,"vote_count":6171},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1584.5115853041812,"type":"confidence_interval","upper":1602.5542657451047}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1593.532925524643,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-719698bd14b21e85e673","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4df50c225e9e5cb8fba5d431","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1582.0912797071308,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f8c6893005929dad7dcce46742401ccb807215a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1998,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=10","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1998,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1582.0912797071308,"rating_lower":1574.3209971204221,"rating_upper":1589.8615622938394,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.717281974247038,"vote_count":8775},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1574.3209971204221,"type":"confidence_interval","upper":1589.8615622938394}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1582.0912797071308,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1785fd11a3878e8d0887","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-385947df459796b0dfbff214","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1582.0912797071308,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f8c6893005929dad7dcce46742401ccb807215a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2160,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=172","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2160,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1582.0912797071308,"rating_lower":1574.3209971204221,"rating_upper":1589.8615622938394,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.717281974247038,"vote_count":8775},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1574.3209971204221,"type":"confidence_interval","upper":1589.8615622938394}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1582.0912797071308,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a626edee88c2c86a0395","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-588f2dab6a6a9937fc25e229","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.2","modelRef":"glm-5.2-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1536.5926280216395,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f8c6893005929dad7dcce46742401ccb807215a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2289,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=301","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2289,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1536.5926280216395,"rating_lower":1516.5918927763114,"rating_upper":1556.593363266968,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":104.13476468854465,"vote_count":999},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1516.5918927763114,"type":"confidence_interval","upper":1556.593363266968}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1536.5926280216395,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9bf4931e38a264254df7","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a56ec55da46855531df8d1fa","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1598.6001848847525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"509a4a14030b725a362561601e8f5f31a1d7d6cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1995,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=7","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1995,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1598.6001848847525,"rating_lower":1583.4690517589133,"rating_upper":1613.7313180105916,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":59.60006350673301,"vote_count":1916},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1583.4690517589133,"type":"confidence_interval","upper":1613.7313180105916}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1598.6001848847525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-81468b7c14df5686d91f","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8dd04cb44bd39ecefe09c5f6","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1609.2520822615816,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"509a4a14030b725a362561601e8f5f31a1d7d6cf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2392,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=404","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2392,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1609.2520822615816,"rating_lower":1590.0397365583285,"rating_upper":1628.4644279648344,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":96.08699315813789,"vote_count":1280},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1590.0397365583285,"type":"confidence_interval","upper":1628.4644279648344}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1609.2520822615816,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1721436f831817911ab7","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0135e6180c96d8f8ff4ac530","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1598.6001848847525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"509a4a14030b725a362561601e8f5f31a1d7d6cf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2157,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=169","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2157,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1598.6001848847525,"rating_lower":1583.4690517589133,"rating_upper":1613.7313180105916,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":59.60006350673301,"vote_count":1916},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1583.4690517589133,"type":"confidence_interval","upper":1613.7313180105916}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1598.6001848847525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2f9bda0d9e2761726e6b","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-da29e3b24f31ecbaabf2a117","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5.3","modelRef":"glm-5.3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":10,"rawValue":1577.091365387705,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"509a4a14030b725a362561601e8f5f31a1d7d6cf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2277,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=289","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2277,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"zai","rating":1577.091365387705,"rating_lower":1532.4605488963848,"rating_upper":1621.7221818790254,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":518.529515389153,"vote_count":200},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1532.4605488963848,"type":"confidence_interval","upper":1621.7221818790254}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1577.091365387705,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-889b0bb07eafd5e80157","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4ebd14b38ef701842c65229f","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1423.425383722039,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e42b002026e2cc63a327c64e1d8a95922cf850a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2141,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=153","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2141,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1423.425383722039,"rating_lower":1406.8344498127276,"rating_upper":1440.0163176313504,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":71.65483240385204,"vote_count":1445},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1406.8344498127276,"type":"confidence_interval","upper":1440.0163176313504}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1423.425383722039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-14e8bb3493294aad96f6","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dc3b6ecb20f7edc170895922","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":65,"rawValue":1399.8264781078112,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e42b002026e2cc63a327c64e1d8a95922cf850a3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2214,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=226","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2214,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1399.8264781078112,"rating_lower":1385.8242770823067,"rating_upper":1413.828679133316,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.038327549536774,"vote_count":2183},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1385.8242770823067,"type":"confidence_interval","upper":1413.828679133316}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1399.8264781078112,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ed1f28cad3d1a9e748d4","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-23f775f249091f24b1a8a9c5","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":65,"rawValue":1399.8264781078112,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e42b002026e2cc63a327c64e1d8a95922cf850a3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2052,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=64","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2052,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1399.8264781078112,"rating_lower":1385.8242770823067,"rating_upper":1413.828679133316,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.038327549536774,"vote_count":2183},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1385.8242770823067,"type":"confidence_interval","upper":1413.828679133316}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1399.8264781078112,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-27aa2aa90341808be878","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e37a460a58467a7edc782a56","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1383.83048846566,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e42b002026e2cc63a327c64e1d8a95922cf850a3","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2453,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=465","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2453,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1383.83048846566,"rating_lower":1367.122749600145,"rating_upper":1400.538227331175,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":72.66732536463851,"vote_count":1562},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1367.122749600145,"type":"confidence_interval","upper":1400.538227331175}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1383.83048846566,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8b73e4c4cf5e455d11d1","canonicalModelId":"zai/glm-5v-turbo@2026-04-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-94b0b56be1c6149172902ce7","mappingCandidates":["zai/glm-5v-turbo@2026-04-01"],"mappingEvidence":{"confidence":"high","note":"Arena exact GLM-5V-Turbo endpoint spelling.","registry":"model_aliases.json"},"mappingNote":"Arena exact GLM-5V-Turbo endpoint spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GLM-5V-Turbo","modelRef":"glm-5v-turbo","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":83,"rawValue":1364.6644015609772,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e42b002026e2cc63a327c64e1d8a95922cf850a3","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2350,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=362","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2350,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"zai","rating":1364.6644015609772,"rating_lower":1318.8132774435257,"rating_upper":1410.515525678429,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":547.2727109577795,"vote_count":219},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1318.8132774435257,"type":"confidence_interval","upper":1410.515525678429}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1364.6644015609772,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6387c6d47a226da18df7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e945f9d4bf903ea8ee5a2eb8","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":56,"rawValue":1428.9187308641385,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a2e8c70accca01ca15526cd9344f1e462909ca7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2323,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=335","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2323,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1428.9187308641385,"rating_lower":1412.1393565313438,"rating_upper":1445.698105196933,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":73.29179255634246,"vote_count":3037},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1412.1393565313438,"type":"confidence_interval","upper":1445.698105196933}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1428.9187308641385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0e135ab4ae8bb16219f3","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fd2a24dbdd47639de0c2630a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1419.1398643855996,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a2e8c70accca01ca15526cd9344f1e462909ca7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2208,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=220","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2208,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1419.1398643855996,"rating_lower":1403.3939693390937,"rating_upper":1434.8857594321055,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.54142095184027,"vote_count":3037},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1403.3939693390937,"type":"confidence_interval","upper":1434.8857594321055}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1419.1398643855996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-efc82fe911a0eb6935c0","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e29d675a1a84941719db9aa6","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5","modelRef":"gpt-5-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1419.1398643855996,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a2e8c70accca01ca15526cd9344f1e462909ca7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2046,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=58","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2046,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1419.1398643855996,"rating_lower":1403.3939693390937,"rating_upper":1434.8857594321055,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.54142095184027,"vote_count":3037},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1403.3939693390937,"type":"confidence_interval","upper":1434.8857594321055}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1419.1398643855996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b2aa30ef5c368bda3cb1","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a447a837ce9a5ae68227adb1","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1344.786912963157,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2147,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=159","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2147,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1344.786912963157,"rating_lower":1323.324086264749,"rating_upper":1366.249739661565,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":119.91614420134574,"vote_count":1051},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1323.324086264749,"type":"confidence_interval","upper":1366.249739661565}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1344.786912963157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fe71c27becdb6f17fcb2","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a7baf1854bb47aaa76e8976e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":84,"rawValue":1364.1835378758472,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2351,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=363","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2351,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1364.1835378758472,"rating_lower":1352.8464815806901,"rating_upper":1375.520594171004,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":33.45834263461689,"vote_count":7805},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1352.8464815806901,"type":"confidence_interval","upper":1375.520594171004}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1364.1835378758472,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8320b1f36eefe1299889","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c2d9209478d5e960aebbc7f1","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":89,"rawValue":1340.632179382849,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2238,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=250","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2238,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1340.632179382849,"rating_lower":1332.1045547318872,"rating_upper":1349.1598040338108,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.930407842964147,"vote_count":10343},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1332.1045547318872,"type":"confidence_interval","upper":1349.1598040338108}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1340.632179382849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c922acff8885c31768eb","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9e828425f309265621f21d92","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":89,"rawValue":1340.632179382849,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2076,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=88","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2076,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1340.632179382849,"rating_lower":1332.1045547318872,"rating_upper":1349.1598040338108,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.930407842964147,"vote_count":10343},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1332.1045547318872,"type":"confidence_interval","upper":1349.1598040338108}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1340.632179382849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-21de0515f1a00f355b95","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aeae0929a2c3b98903a556ae","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":90,"rawValue":1304.5400830339133,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2474,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=486","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2474,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1304.5400830339133,"rating_lower":1290.6097902637441,"rating_upper":1318.4703758040826,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":50.515459287824115,"vote_count":2533},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1290.6097902637441,"type":"confidence_interval","upper":1318.4703758040826}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1304.5400830339133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-50048ff9239cb0a470d9","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d71dcd0d864d01e84e247583","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":91,"rawValue":1345.8733070042335,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5e158c5a8095ae67d897a6441d1696efdad3895","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2358,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=370","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2358,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1345.8733070042335,"rating_lower":1332.8784286915975,"rating_upper":1358.86818531687,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.95904531125856,"vote_count":4857},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1332.8784286915975,"type":"confidence_interval","upper":1358.86818531687}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1345.8733070042335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-053de8b2f4e3c49585fe","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bea2af75747d983eabdb0adf","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":92,"rawValue":1336.2778561874031,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5e158c5a8095ae67d897a6441d1696efdad3895","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2241,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=253","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2241,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1336.2778561874031,"rating_lower":1324.5903979003376,"rating_upper":1347.9653144744689,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.55854366472617,"vote_count":4873},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1324.5903979003376,"type":"confidence_interval","upper":1347.9653144744689}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1336.2778561874031,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4bc4a8b2812ef4254d52","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b04d885cf2ae5c3ce0ece77e","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":92,"rawValue":1336.2778561874031,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5e158c5a8095ae67d897a6441d1696efdad3895","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2079,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=91","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2079,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1336.2778561874031,"rating_lower":1324.5903979003376,"rating_upper":1347.9653144744689,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.55854366472617,"vote_count":4873},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1324.5903979003376,"type":"confidence_interval","upper":1347.9653144744689}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1336.2778561874031,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-25dadda5a2630df94665","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-58a5f67cc025d2afb3b708e6","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":106,"rawValue":1253.6184669539425,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5e29ddd5725edf42f0c5751a924106bd7379d75","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2373,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=385","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2373,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1253.6184669539425,"rating_lower":1230.580344307717,"rating_upper":1276.656589600168,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":138.16498362635608,"vote_count":1133},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1230.580344307717,"type":"confidence_interval","upper":1276.656589600168}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1253.6184669539425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-98a6f9748f16bf35471c","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-318ce3bc3af113b875505a49","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":106,"rawValue":1243.7391069949672,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5e29ddd5725edf42f0c5751a924106bd7379d75","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2255,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=267","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2255,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1243.7391069949672,"rating_lower":1221.4455117027103,"rating_upper":1266.032702287224,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":129.37907556825814,"vote_count":1133},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.4455117027103,"type":"confidence_interval","upper":1266.032702287224}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1243.7391069949672,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4be514f8b8ed2770557c","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-070ad373581da8f294004efb","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":106,"rawValue":1243.7391069949672,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5e29ddd5725edf42f0c5751a924106bd7379d75","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2093,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=105","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2093,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1243.7391069949672,"rating_lower":1221.4455117027103,"rating_upper":1266.032702287224,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":129.37907556825814,"vote_count":1133},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1221.4455117027103,"type":"confidence_interval","upper":1266.032702287224}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1243.7391069949672,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-838bd3f5bddd353cdd4d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b4bde459b2d610572af56674","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-high","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1419.758619176206,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2142,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=154","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2142,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1419.758619176206,"rating_lower":1397.8029615917296,"rating_upper":1441.714276760682,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":125.48641609014182,"vote_count":947},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1397.8029615917296,"type":"confidence_interval","upper":1441.714276760682}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1419.758619176206,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f6b13fbf779c1c961cb0","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fca811f54e2c3ed9468f8413","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1401.0621185609332,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2335,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=347","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2335,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1401.0621185609332,"rating_lower":1388.1065719135986,"rating_upper":1414.0176652082678,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.693345878672396,"vote_count":4737},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1388.1065719135986,"type":"confidence_interval","upper":1414.0176652082678}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1401.0621185609332,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-91f4ecc246ede7c6e663","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-919b22e39cedf40b080f8604","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":70,"rawValue":1391.014390336067,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2219,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=231","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2219,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1391.014390336067,"rating_lower":1379.403605309813,"rating_upper":1402.6251753623208,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.09352441828876,"vote_count":4758},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1379.403605309813,"type":"confidence_interval","upper":1402.6251753623208}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1391.014390336067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-130e03f9c53ef2a757d6","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3ea0740e643f3b8782441220","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.1","modelRef":"gpt-5.1-medium","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":70,"rawValue":1391.014390336067,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"609e3b27a64debd63fe5fe89eacfccd097ef9711","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2057,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=69","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2057,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1391.014390336067,"rating_lower":1379.403605309813,"rating_upper":1402.6251753623208,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.09352441828876,"vote_count":4758},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1379.403605309813,"type":"confidence_interval","upper":1402.6251753623208}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1391.014390336067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d58e70c382cf571605cf","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-370c04f673537a9fa15c7231","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1428.395452781471,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc0df44fba215a1d6557910996af01a5eac791c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2324,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=336","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2324,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1428.395452781471,"rating_lower":1404.7029527795669,"rating_upper":1452.0879527833754,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":146.12536084372306,"vote_count":1057},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1404.7029527795669,"type":"confidence_interval","upper":1452.0879527833754}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1428.395452781471,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-00bb75d586499f6ab53a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b6264b7f3d3887dbfebaa362","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":60,"rawValue":1417.810641177648,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc0df44fba215a1d6557910996af01a5eac791c3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2209,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=221","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2209,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1417.810641177648,"rating_lower":1394.9142485295451,"rating_upper":1440.707033825751,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":136.4702371588554,"vote_count":1057},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.9142485295451,"type":"confidence_interval","upper":1440.707033825751}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1417.810641177648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b68978394f90c3e0a99a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-23f08737fb6656b29d6d09c0","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2","modelRef":"gpt-5.2","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":60,"rawValue":1417.810641177648,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc0df44fba215a1d6557910996af01a5eac791c3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2047,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=59","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2047,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1417.810641177648,"rating_lower":1394.9142485295451,"rating_upper":1440.707033825751,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":136.4702371588554,"vote_count":1057},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.9142485295451,"type":"confidence_interval","upper":1440.707033825751}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1417.810641177648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-336b353db83058ece048","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9c31d3be721ab51e6aaf1361","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":86,"rawValue":1329.643790401107,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba5d0fcbd409e78f06a1185a4d19460eff0ef9b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2470,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=482","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2470,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1329.643790401107,"rating_lower":1318.9677888074289,"rating_upper":1340.3197919947854,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.67024126725505,"vote_count":4136},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1318.9677888074289,"type":"confidence_interval","upper":1340.3197919947854}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1329.643790401107,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fb7f21d1083aa31bba6a","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dfd720a95b5233bf466b00ac","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":90,"rawValue":1348.245079097252,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba5d0fcbd409e78f06a1185a4d19460eff0ef9b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2357,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=369","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2357,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1348.245079097252,"rating_lower":1333.0478796743587,"rating_upper":1363.4422785201455,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":60.121657182690036,"vote_count":2284},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1333.0478796743587,"type":"confidence_interval","upper":1363.4422785201455}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1348.245079097252,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1d15a9cdf66fbcbe8071","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e948f5f2699187e5a76a83b5","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":91,"rawValue":1338.2841667454431,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba5d0fcbd409e78f06a1185a4d19460eff0ef9b7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2078,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=90","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2078,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1338.2841667454431,"rating_lower":1329.2862623423914,"rating_upper":1347.2820711484947,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.075921264678758,"vote_count":6424},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1329.2862623423914,"type":"confidence_interval","upper":1347.2820711484947}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1338.2841667454431,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-61178a72439aa5158e96","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-08721cf40908c35093d1eaf9","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":91,"rawValue":1338.2841667454431,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba5d0fcbd409e78f06a1185a4d19460eff0ef9b7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2240,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=252","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2240,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1338.2841667454431,"rating_lower":1329.2862623423914,"rating_upper":1347.2820711484947,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.075921264678758,"vote_count":6424},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1329.2862623423914,"type":"confidence_interval","upper":1347.2820711484947}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1338.2841667454431,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1ccf2b4a8ac56e47f665","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-70543969743a3d4f5e96468f","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1444.2663125087988,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2136,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=148","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2136,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1444.2663125087988,"rating_lower":1430.03432367796,"rating_upper":1458.4983013396375,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":52.727236067187256,"vote_count":2539},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.03432367796,"type":"confidence_interval","upper":1458.4983013396375}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1444.2663125087988,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b0923bedcc3a47ce68da","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c744b5db5d4b49c44e8b8ad5","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1427.3925039892774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2326,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=338","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2326,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1427.3925039892774,"rating_lower":1394.829202452992,"rating_upper":1459.9558055255627,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":276.03279286264546,"vote_count":373},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.829202452992,"type":"confidence_interval","upper":1459.9558055255627}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1427.3925039892774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3186df321a447c6e3133","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0aa3f555043ef375fe1ffb4c","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":59,"rawValue":1397.9793187551898,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2443,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=455","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2443,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1397.9793187551898,"rating_lower":1383.1541314729905,"rating_upper":1412.8045060373893,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.21424807895577,"vote_count":2249},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1383.1541314729905,"type":"confidence_interval","upper":1412.8045060373893}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1397.9793187551898,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c7e395669ad261ee1aa5","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-43d76f43745ea44f0d28750b","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":61,"rawValue":1408.3779995790048,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2048,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=60","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2048,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1408.3779995790048,"rating_lower":1394.5125899953023,"rating_upper":1422.2434091627074,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":50.045983022952406,"vote_count":2526},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.5125899953023,"type":"confidence_interval","upper":1422.2434091627074}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1408.3779995790048,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f10b97d52dc33fe2c0a5","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3dd017d866ecf478ff6958af","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":61,"rawValue":1408.3779995790048,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2210,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=222","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2210,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1408.3779995790048,"rating_lower":1394.5125899953023,"rating_upper":1422.2434091627074,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":50.045983022952406,"vote_count":2526},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1394.5125899953023,"type":"confidence_interval","upper":1422.2434091627074}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1408.3779995790048,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bad8a3d15c7a824b0bee","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b35e8499a2354becb5dccbd8","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1419.8166295772658,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2330,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=342","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2330,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1419.8166295772658,"rating_lower":1381.890933904001,"rating_upper":1457.742325250531,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":374.43025148482747,"vote_count":271},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1381.890933904001,"type":"confidence_interval","upper":1457.742325250531}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1419.8166295772658,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-09433cb8cd6296e9d431","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-01da5ef65e22d8df28a9d1e2","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":76,"rawValue":1354.5204484579363,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2460,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=472","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2460,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1354.5204484579363,"rating_lower":1342.2906344607811,"rating_upper":1366.7502624550914,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.93529968335933,"vote_count":3106},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1342.2906344607811,"type":"confidence_interval","upper":1366.7502624550914}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1354.5204484579363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-73de8ff59231b1ca408a","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aa944e8f0f8978b69e188c31","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":79,"rawValue":1368.9959519462375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2228,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=240","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2228,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1368.9959519462375,"rating_lower":1357.48948940762,"rating_upper":1380.5024144848549,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.46572938368264,"vote_count":3504},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1357.48948940762,"type":"confidence_interval","upper":1380.5024144848549}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1368.9959519462375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e4057fd9520e07180b58","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-691f65be399bbc60bccc7fd9","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":79,"rawValue":1368.9959519462375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f753bb061d0a561a9d7aa20792bf5f383b1b869b","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2066,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=78","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2066,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1368.9959519462375,"rating_lower":1357.48948940762,"rating_upper":1380.5024144848549,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.46572938368264,"vote_count":3504},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1357.48948940762,"type":"confidence_interval","upper":1380.5024144848549}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1368.9959519462375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e0d6b2bf28901a6d5fa1","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dfe2f597e7e151bffc42bb40","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1442.2774568784275,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2138,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=150","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2138,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1442.2774568784275,"rating_lower":1428.2411932877626,"rating_upper":1456.3137204690925,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":51.286947168426565,"vote_count":2180},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1428.2411932877626,"type":"confidence_interval","upper":1456.3137204690925}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1442.2774568784275,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f453a7f8e90d47d37e44","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8de2a382ecaaa115f7074d16","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":60,"rawValue":1423.2094096779504,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2327,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=339","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2327,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1423.2094096779504,"rating_lower":1381.0286532466096,"rating_upper":1465.3901661092914,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":463.16159984206547,"vote_count":203},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1381.0286532466096,"type":"confidence_interval","upper":1465.3901661092914}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1423.2094096779504,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5e7e48b2d0d1cb09d2c4","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6f73130f09be67dec1cc7447","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1388.0096165628515,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2447,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=459","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2447,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1388.0096165628515,"rating_lower":1371.3220987154032,"rating_upper":1404.6971344102997,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":72.49153639465186,"vote_count":1477},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1371.3220987154032,"type":"confidence_interval","upper":1404.6971344102997}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1388.0096165628515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4f261494a0231f638a51","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c50406e4e7590958a88e2abd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":71,"rawValue":1390.556992986316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2220,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=232","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2220,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1390.556992986316,"rating_lower":1376.24180671023,"rating_upper":1404.8721792624015,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":53.34550432119484,"vote_count":1951},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1376.24180671023,"type":"confidence_interval","upper":1404.8721792624015}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1390.556992986316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-76fba54d383513bb8f3c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b8cca007d58df9397968e84f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":71,"rawValue":1390.556992986316,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2058,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=70","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2058,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1390.556992986316,"rating_lower":1376.24180671023,"rating_upper":1404.8721792624015,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":53.34550432119484,"vote_count":1951},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1376.24180671023,"type":"confidence_interval","upper":1404.8721792624015}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1390.556992986316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6e82bc26572257535917","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d6e0ee0bb9196c45004bafde","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":42,"rawValue":1476.7572112847733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2309,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=321","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2309,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1476.7572112847733,"rating_lower":1415.8209514162259,"rating_upper":1537.6934711533208,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":966.6191777883491,"vote_count":113},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1415.8209514162259,"type":"confidence_interval","upper":1537.6934711533208}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1476.7572112847733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-405686fd5d84d2818bc5","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-807795cd40c2e0d071e03057","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1462.8963395174703,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2193,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=205","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2193,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1462.8963395174703,"rating_lower":1443.4568331170024,"rating_upper":1482.3358459179383,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":98.37263048561168,"vote_count":1279},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1443.4568331170024,"type":"confidence_interval","upper":1482.3358459179383}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1462.8963395174703,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-990a51f2f17864a19b1f","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0ed268695e6e009974f5b9f0","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1462.8963395174703,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2031,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=43","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2031,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1462.8963395174703,"rating_lower":1443.4568331170024,"rating_upper":1482.3358459179383,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":98.37263048561168,"vote_count":1279},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1443.4568331170024,"type":"confidence_interval","upper":1482.3358459179383}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1462.8963395174703,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ccfd8f7375db73754044","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4023cc7643c79283cbc32e3a","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1453.1340210108017,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2429,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=441","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2429,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1453.1340210108017,"rating_lower":1432.793651012502,"rating_upper":1473.474391009101,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":107.70144129592997,"vote_count":1166},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1432.793651012502,"type":"confidence_interval","upper":1473.474391009101}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1453.1340210108017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d156abe77182aa70820a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6bfcb5a64d0adac3ca4c5ef9","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-medium (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":43,"rawValue":1474.048924083781,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2310,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=322","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2310,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1474.048924083781,"rating_lower":1420.572906033768,"rating_upper":1527.524942133794,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":744.4266982845271,"vote_count":130},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1420.572906033768,"type":"confidence_interval","upper":1527.524942133794}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1474.048924083781,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e5c1f06cb77539affdca","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a52c8de424ab3b34711a7d73","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-medium (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1429.1866596212149,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2433,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=445","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2433,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1429.1866596212149,"rating_lower":1409.4433037681458,"rating_upper":1448.9300154742841,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":101.47189350073094,"vote_count":1188},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1409.4433037681458,"type":"confidence_interval","upper":1448.9300154742841}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1429.1866596212149,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7eb28f2f2427540cc712","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-896ee6e64ca816533abe73b7","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-medium (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1442.1566799903471,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2037,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=49","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2037,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1442.1566799903471,"rating_lower":1423.4927850294878,"rating_upper":1460.8205749512063,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":90.6793464070111,"vote_count":1318},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1423.4927850294878,"type":"confidence_interval","upper":1460.8205749512063}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1442.1566799903471,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":13,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3a642f11a7315db8376a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-49f6eb884915dbabd2d7d6cb","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","registry":"model_aliases.json"},"mappingNote":"Arena harness presentations of GPT-5.4; harness and effort remain visible on each public row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4","modelRef":"gpt-5.4-medium (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":50,"rawValue":1442.1566799903471,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1df693b72c7039640daff93c15d3c674aea42a45","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2199,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=211","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2199,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1442.1566799903471,"rating_lower":1423.4927850294878,"rating_upper":1460.8205749512063,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":90.6793464070111,"vote_count":1318},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1423.4927850294878,"type":"confidence_interval","upper":1460.8205749512063}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1442.1566799903471,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4eadae139c68a83b4e79","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-620403d76fe410047ba04822","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":62,"rawValue":1422.8173553543616,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"439b6b04087f730d232f73adf8fb9a7bce2f7a02","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2329,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=341","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2329,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1422.8173553543616,"rating_lower":1405.483730535026,"rating_upper":1440.1509801736968,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":78.21365876914143,"vote_count":1340},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1405.483730535026,"type":"confidence_interval","upper":1440.1509801736968}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1422.8173553543616,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-014749a93bf5001bd495","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5d968d78c74a760d88ee1dc7","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":66,"rawValue":1387.0361208776533,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"439b6b04087f730d232f73adf8fb9a7bce2f7a02","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2450,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=462","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2450,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1387.0361208776533,"rating_lower":1379.7756071441972,"rating_upper":1394.2966346111095,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.722666865443662,"vote_count":9331},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1379.7756071441972,"type":"confidence_interval","upper":1394.2966346111095}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1387.0361208776533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf41c00e109ea9e68ed2","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9a49a5e8f65753fbf06257fe","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1397.0815445066285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"439b6b04087f730d232f73adf8fb9a7bce2f7a02","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2217,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=229","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2217,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1397.0815445066285,"rating_lower":1390.392620307159,"rating_upper":1403.7704687060977,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.64706145103569,"vote_count":11294},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1390.392620307159,"type":"confidence_interval","upper":1403.7704687060977}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1397.0815445066285,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-65a6ae856aa2873fc442","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-52aae1e04e7df00ffb027ea2","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":68,"rawValue":1397.0815445066285,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"439b6b04087f730d232f73adf8fb9a7bce2f7a02","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2055,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=67","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2055,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1397.0815445066285,"rating_lower":1390.392620307159,"rating_upper":1403.7704687060977,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.64706145103569,"vote_count":11294},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1390.392620307159,"type":"confidence_interval","upper":1403.7704687060977}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1397.0815445066285,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2ad3ec73adbd470938d3","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f66bcb3b3686fda44f6314c7","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5 (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":22,"rawValue":1500.599108273734,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2127,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=139","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2127,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1500.599108273734,"rating_lower":1491.1074961458075,"rating_upper":1510.0907204016603,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":23.45221047318776,"vote_count":6115},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1491.1074961458075,"type":"confidence_interval","upper":1510.0907204016603}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1500.599108273734,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cea61fc5ac0e97588360","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cdc2f48280c8dcbca4e90d0c","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5 (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1501.1631416341384,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2302,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=314","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2302,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1501.1631416341384,"rating_lower":1485.60050215415,"rating_upper":1516.7257811141271,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":63.04785730862914,"vote_count":1672},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1485.60050215415,"type":"confidence_interval","upper":1516.7257811141271}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1501.1631416341384,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-17323361916b7d0d145d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fb2bc522311bb323a8fc996f","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5 (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1457.5140563377497,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":13,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2195,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=207","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2195,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1457.5140563377497,"rating_lower":1451.2079719308506,"rating_upper":1463.8201407446488,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.351978871336797,"vote_count":13672},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1451.2079719308506,"type":"confidence_interval","upper":1463.8201407446488}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1457.5140563377497,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f50292c067fa43a2f331","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-da0b9700dd8b122b7608eaa9","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5 (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1447.234029142822,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":14,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2430,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=442","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2430,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1447.234029142822,"rating_lower":1440.2492799424947,"rating_upper":1454.218778343149,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.700050597615737,"vote_count":10620},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1440.2492799424947,"type":"confidence_interval","upper":1454.218778343149}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1447.234029142822,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-701713b2643a2d5b6cd0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-85e10562a9dafd9f278b5211","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5 (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1457.5140563377497,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":15,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2033,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=45","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2033,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1457.5140563377497,"rating_lower":1451.2079719308506,"rating_upper":1463.8201407446488,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.351978871336797,"vote_count":13672},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1451.2079719308506,"type":"confidence_interval","upper":1463.8201407446488}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1457.5140563377497,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c1397956ca87be9f6892","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-983b734f32e60970553b4710","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":19,"rawValue":1516.399261218979,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2124,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=136","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2124,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1516.399261218979,"rating_lower":1507.1112935542726,"rating_upper":1525.6872288836853,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":22.45666226484275,"vote_count":6477},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1507.1112935542726,"type":"confidence_interval","upper":1525.6872288836853}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1516.399261218979,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-75b49066ac5a6fe90bf1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7f3a5754a6859bda3cff786d","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1539.9274492312652,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2287,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=299","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2287,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1539.9274492312652,"rating_lower":1524.5499037632958,"rating_upper":1555.3049946992342,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":61.55705830961712,"vote_count":1696},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1524.5499037632958,"type":"confidence_interval","upper":1555.3049946992342}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1539.9274492312652,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8a23b61c48507b00ddcd","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-700c9b6de962d6b94ca2b42a","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1486.1697842138751,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2025,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=37","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2025,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1486.1697842138751,"rating_lower":1479.897137839951,"rating_upper":1492.442430587799,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.242487130238562,"vote_count":15630},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1479.897137839951,"type":"confidence_interval","upper":1492.442430587799}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1486.1697842138751,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-edf8f6d6edbfb6c4e39f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-670e582fc746333379ca5b30","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1476.9618265794325,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2422,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=434","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2422,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1476.9618265794325,"rating_lower":1470.0319984813834,"rating_upper":1483.8916546774817,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.501114735321634,"vote_count":10995},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1470.0319984813834,"type":"confidence_interval","upper":1483.8916546774817}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1476.9618265794325,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a6f91218a101920a92f5","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2d008bb685e31b333fcea4b3","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-high (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1486.1697842138751,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2187,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=199","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2187,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1486.1697842138751,"rating_lower":1479.897137839951,"rating_upper":1492.442430587799,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.242487130238562,"vote_count":15630},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1479.897137839951,"type":"confidence_interval","upper":1492.442430587799}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1486.1697842138751,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cbfe62ac24f4aabd1707","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-653e81269db81930b92a16da","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":16,"rawValue":1529.7172814531418,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2121,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=133","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2121,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1529.7172814531418,"rating_lower":1520.0410702286335,"rating_upper":1539.3934926776496,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.373309211831206,"vote_count":5971},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1520.0410702286335,"type":"confidence_interval","upper":1539.3934926776496}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1529.7172814531418,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5862d31106a8e3cfc6cf","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-69d78251118af24ba7924f68","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1544.6596061546925,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2285,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=297","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2285,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1544.6596061546925,"rating_lower":1528.4290549702841,"rating_upper":1560.8901573391013,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":68.57571668622013,"vote_count":1614},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1528.4290549702841,"type":"confidence_interval","upper":1560.8901573391013}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1544.6596061546925,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-70ffd326896688f00b9c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-346aebad61cad329dc8ad9dd","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1500.3687592412523,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2417,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=429","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2417,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1500.3687592412523,"rating_lower":1493.3894106470561,"rating_upper":1507.3481078354484,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.680418839034859,"vote_count":10704},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1493.3894106470561,"type":"confidence_interval","upper":1507.3481078354484}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1500.3687592412523,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-3077abd8851b2550ef70","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3172a18edae7f50e863721b0","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1508.3502120258615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2020,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=32","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2020,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1508.3502120258615,"rating_lower":1502.0037464065758,"rating_upper":1514.696677645147,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.484981809461926,"vote_count":13648},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1502.0037464065758,"type":"confidence_interval","upper":1514.696677645147}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1508.3502120258615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":15,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4dd1b20c66014230f418","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-14a6e7eaf59d01fb4166112f","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Arena exposes the same GPT-5.5 release through a harness or search presentation; those settings are retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":33,"rawValue":1508.3502120258615,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3584ff83927c01335442cbe85c02d4c47a37e2","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2182,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=194","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2182,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1508.3502120258615,"rating_lower":1502.0037464065758,"rating_upper":1514.696677645147,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.484981809461926,"vote_count":13648},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1502.0037464065758,"type":"confidence_interval","upper":1514.696677645147}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1508.3502120258615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7c3fd1722d9a64e3724b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c872900b2bc23f73678d2066","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":24,"rawValue":1495.8214394408,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04ec0a23b143f1b30a8be6bb53b391285d246f4a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2129,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=141","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2129,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1495.8214394408,"rating_lower":1481.0091883676591,"rating_upper":1510.6336905139406,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.11444326093707,"vote_count":1693},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1481.0091883676591,"type":"confidence_interval","upper":1510.6336905139406}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1495.8214394408,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2e6778375d0ebd3f8eeb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b26b0de0090e68c2d35c3512","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":26,"rawValue":1532.8442516988694,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04ec0a23b143f1b30a8be6bb53b391285d246f4a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2293,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=305","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2293,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1532.8442516988694,"rating_lower":1507.2430389574815,"rating_upper":1558.4454644402572,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":170.61801894087878,"vote_count":629},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1507.2430389574815,"type":"confidence_interval","upper":1558.4454644402572}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1532.8442516988694,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cd03ce0d17a7b68dafe4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-90fcce7acbfc8fd771857b06","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1517.597333795525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04ec0a23b143f1b30a8be6bb53b391285d246f4a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2178,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=190","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2178,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1517.597333795525,"rating_lower":1508.8851683785506,"rating_upper":1526.309499212499,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.758594272528654,"vote_count":5742},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1508.8851683785506,"type":"confidence_interval","upper":1526.309499212499}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1517.597333795525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-aa89180f2b4dd35dd294","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8a3246e5f093f587bfedb30c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1517.597333795525,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04ec0a23b143f1b30a8be6bb53b391285d246f4a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2016,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=28","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2016,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1517.597333795525,"rating_lower":1508.8851683785506,"rating_upper":1526.309499212499,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.758594272528654,"vote_count":5742},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1508.8851683785506,"type":"confidence_interval","upper":1526.309499212499}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1517.597333795525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1bbe7a5061c91bb07136","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2e5f54e52e5ec86ee115db71","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1505.9475142861716,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04ec0a23b143f1b30a8be6bb53b391285d246f4a","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2414,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=426","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2414,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1505.9475142861716,"rating_lower":1495.4047513932699,"rating_upper":1516.4902771790735,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.93428111663665,"vote_count":3832},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1495.4047513932699,"type":"confidence_interval","upper":1516.4902771790735}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1505.9475142861716,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-01721ad9bd2c36c2db57","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-403eeff24ad4c71328d0181f","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1606.4791387238115,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d75159f163a8faea8432c67c50523ac8831bb38","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2109,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=121","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2109,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1606.4791387238115,"rating_lower":1591.5424260723612,"rating_upper":1621.4158513752614,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.07829661744892,"vote_count":1758},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1591.5424260723612,"type":"confidence_interval","upper":1621.4158513752614}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1606.4791387238115,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6a18a1d30e5b21990f4a","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-44a2ea67a334bb2ded113eed","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1624.9415689051705,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d75159f163a8faea8432c67c50523ac8831bb38","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2391,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=403","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2391,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1624.9415689051705,"rating_lower":1615.9165016044442,"rating_upper":1633.9666362058967,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.203361427136535,"vote_count":6475},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1615.9165016044442,"type":"confidence_interval","upper":1633.9666362058967}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1624.9415689051705,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-124db8e9760a774e8419","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-434d8c41c52dc4e70b6aa4e0","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1618.7118275567602,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d75159f163a8faea8432c67c50523ac8831bb38","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1994,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=6","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1994,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1618.7118275567602,"rating_lower":1611.0282579573814,"rating_upper":1626.3953971561389,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.368443225386725,"vote_count":9499},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1611.0282579573814,"type":"confidence_interval","upper":1626.3953971561389}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1618.7118275567602,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-37c45ff4e29537087ac3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-42e7d3050c134f035cc4714f","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1618.7118275567602,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d75159f163a8faea8432c67c50523ac8831bb38","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2156,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=168","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2156,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1618.7118275567602,"rating_lower":1611.0282579573814,"rating_upper":1626.3953971561389,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.368443225386725,"vote_count":9499},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1611.0282579573814,"type":"confidence_interval","upper":1626.3953971561389}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1618.7118275567602,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0ed1aaad9c5be629aee9","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3caca9e58b0cab5bd1abbec1","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1618.2748501551378,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d75159f163a8faea8432c67c50523ac8831bb38","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2275,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=287","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2275,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1618.2748501551378,"rating_lower":1596.8538671215156,"rating_upper":1639.6958331887604,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":119.44902588955252,"vote_count":1016},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1596.8538671215156,"type":"confidence_interval","upper":1639.6958331887604}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1618.2748501551378,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7128ce19faa5691faac4","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d1d392e0c4fccf49f435e55e","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh (codex-harness)","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1534.2537401820946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de6573e4dfb8082b717f36d856c5e6b9db71d72f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2120,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=132","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2120,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1534.2537401820946,"rating_lower":1519.2864182915337,"rating_upper":1549.2210620726555,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.31657581980349,"vote_count":1649},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1519.2864182915337,"type":"confidence_interval","upper":1549.2210620726555}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1534.2537401820946,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-24fac208754f371e1621","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8c7fb33abe8f0fdf65524cc3","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":23,"rawValue":1536.0699760981693,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de6573e4dfb8082b717f36d856c5e6b9db71d72f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2290,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=302","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2290,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1536.0699760981693,"rating_lower":1509.9952773253153,"rating_upper":1562.1446748710227,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":176.98742790953787,"vote_count":586},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1509.9952773253153,"type":"confidence_interval","upper":1562.1446748710227}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1536.0699760981693,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-40fd82236198ca4406df","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3c344f327fc8b742fd636563","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1519.8677186640261,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de6573e4dfb8082b717f36d856c5e6b9db71d72f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2176,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=188","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2176,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1519.8677186640261,"rating_lower":1511.0538755534021,"rating_upper":1528.68156177465,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.222481615631608,"vote_count":5677},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1511.0538755534021,"type":"confidence_interval","upper":1528.68156177465}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1519.8677186640261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ebc4893e28ed8265f5e3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1e5f205634dce202ea758a17","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1519.8677186640261,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de6573e4dfb8082b717f36d856c5e6b9db71d72f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2014,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=26","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2014,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1519.8677186640261,"rating_lower":1511.0538755534021,"rating_upper":1528.68156177465,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.222481615631608,"vote_count":5677},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1511.0538755534021,"type":"confidence_interval","upper":1528.68156177465}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1519.8677186640261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-62abd208a3d5812e81e0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-89bf76c2b3248e5afc3401f9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-xhigh (codex-harness)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":29,"rawValue":1511.7410468620947,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de6573e4dfb8082b717f36d856c5e6b9db71d72f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2413,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=425","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2413,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"openai","rating":1511.7410468620947,"rating_lower":1501.2788330673368,"rating_upper":1522.2032606568525,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":28.49384116720646,"vote_count":3892},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.2788330673368,"type":"confidence_interval","upper":1522.2032606568525}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1511.7410468620947,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c644b30c2fd4bedfa6a1","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b58d15d3d9736b63a681e633","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-reasoning","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":116,"rawValue":1170.2161843158192,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"95ebe934847d4e1cd59325bdc9478a44df06ee3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2383,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=395","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2383,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1170.2161843158192,"rating_lower":1141.924454207494,"rating_upper":1198.5079144241445,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":208.364069454658,"vote_count":733},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1141.924454207494,"type":"confidence_interval","upper":1198.5079144241445}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1170.2161843158192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-17a0baf404b88047e9b7","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ce55eaa595b34dae3fbbc63b","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-reasoning","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":117,"rawValue":1161.8732122243282,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"95ebe934847d4e1cd59325bdc9478a44df06ee3e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2104,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=116","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2104,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1161.8732122243282,"rating_lower":1134.2245201653152,"rating_upper":1189.5219042833414,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":198.99996544437707,"vote_count":733},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1134.2245201653152,"type":"confidence_interval","upper":1189.5219042833414}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1161.8732122243282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2aff0c912afd4a4c0599","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-63692b37dab1e9fea6108c8c","mappingCandidates":["xai/grok-4-fast@2025-09-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4 Fast","modelRef":"grok-4-fast-reasoning","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":117,"rawValue":1161.8732122243282,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"95ebe934847d4e1cd59325bdc9478a44df06ee3e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2266,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=278","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2266,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1161.8732122243282,"rating_lower":1134.2245201653152,"rating_upper":1189.5219042833414,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":198.99996544437707,"vote_count":733},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1134.2245201653152,"type":"confidence_interval","upper":1189.5219042833414}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1161.8732122243282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-79e81ae18c013665dc15","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a7b5b82860d1faa6232c4ae0","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1574.3408563901694,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0be68d8dcaf11a6949a39b8d52ffeaf6369e7dd2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2111,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=123","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2111,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1574.3408563901694,"rating_lower":1559.3069808168402,"rating_upper":1589.374731963499,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.83634975774994,"vote_count":1669},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1559.3069808168402,"type":"confidence_interval","upper":1589.374731963499}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1574.3408563901694,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bad221512e9f7734b92f","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f44471f22c4fbc8003f3d5f2","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":11,"rawValue":1576.772418486707,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0be68d8dcaf11a6949a39b8d52ffeaf6369e7dd2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2278,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=290","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2278,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1576.772418486707,"rating_lower":1551.9433193915984,"rating_upper":1601.6015175818156,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":160.48178326257067,"vote_count":683},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1551.9433193915984,"type":"confidence_interval","upper":1601.6015175818156}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1576.772418486707,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1456ec3029b56ceba485","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-520a5cdafed9d0025772ae38","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":15,"rawValue":1558.6065361696376,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0be68d8dcaf11a6949a39b8d52ffeaf6369e7dd2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2399,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=411","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2399,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1558.6065361696376,"rating_lower":1548.7157978243854,"rating_upper":1568.4972745148898,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.46602985492003,"vote_count":4453},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1548.7157978243854,"type":"confidence_interval","upper":1568.4972745148898}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1558.6065361696376,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6718d10e09f3e90b496b","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ec47e9947e77193f8b51f77d","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1556.3617214249728,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0be68d8dcaf11a6949a39b8d52ffeaf6369e7dd2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2004,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=16","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2004,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1556.3617214249728,"rating_lower":1547.9344155988845,"rating_upper":1564.7890272510613,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.487633683286592,"vote_count":6325},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1547.9344155988845,"type":"confidence_interval","upper":1564.7890272510613}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1556.3617214249728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-183a2edd4ce7fefc49e7","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6de4e8abe5f0896579f9f949","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Arena base spellings of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Arena base spellings of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":17,"rawValue":1556.3617214249728,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0be68d8dcaf11a6949a39b8d52ffeaf6369e7dd2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2166,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=178","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2166,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1556.3617214249728,"rating_lower":1547.9344155988845,"rating_upper":1564.7890272510613,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.487633683286592,"vote_count":6325},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1547.9344155988845,"type":"confidence_interval","upper":1564.7890272510613}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1556.3617214249728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-96f3eec14181d03fe0d0","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d6e9d715972c09846d03dba1","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1638.3158059802884,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a77b4c9cf1d865ed7ad66d298662a92c3b6658f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2389,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=401","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2389,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1638.3158059802884,"rating_lower":1617.2905126009493,"rating_upper":1659.3410993596272,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":115.07684510526569,"vote_count":1064},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1617.2905126009493,"type":"confidence_interval","upper":1659.3410993596272}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1638.3158059802884,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2e1ca4c5b2b3245fcfbc","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-497d9f7fb0c216d733076d12","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1628.8418761621558,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a77b4c9cf1d865ed7ad66d298662a92c3b6658f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1992,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=4","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1992,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1628.8418761621558,"rating_lower":1611.7102596507304,"rating_upper":1645.9734926735814,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":76.4012574372754,"vote_count":1523},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1611.7102596507304,"type":"confidence_interval","upper":1645.9734926735814}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1628.8418761621558,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1d3aa73076be81079558","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-430c25303d80cee6d65ba469","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1628.8418761621558,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a77b4c9cf1d865ed7ad66d298662a92c3b6658f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2154,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=166","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2154,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1628.8418761621558,"rating_lower":1611.7102596507304,"rating_upper":1645.9734926735814,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":76.4012574372754,"vote_count":1523},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1611.7102596507304,"type":"confidence_interval","upper":1645.9734926735814}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1628.8418761621558,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2c7d39781822e73e6ba0","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-314baa0e2dc0300b2ee08546","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Arena high-effort presentation of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Arena high-effort presentation of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Grok 4.6","modelRef":"grok-4.6-high","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":5,"rawValue":1631.454941066038,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a77b4c9cf1d865ed7ad66d298662a92c3b6658f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2272,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=284","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2272,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xai","rating":1631.454941066038,"rating_lower":1580.6226275465679,"rating_upper":1682.2872545855084,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":672.6413631774453,"vote_count":181},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1580.6226275465679,"type":"confidence_interval","upper":1682.2872545855084}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1631.454941066038,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-dbf2dc9bb032313c38cc","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-24552bb92a081c8ba30b3f2d","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":38,"rawValue":1419.3949380571096,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2143,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=155","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2143,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1419.3949380571096,"rating_lower":1396.9701712941755,"rating_upper":1441.8197048200436,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":130.90604060702228,"vote_count":896},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1396.9701712941755,"type":"confidence_interval","upper":1441.8197048200436}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1419.3949380571096,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-43c31a35a9de25617c60","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3a08bb4e9573d63960565635","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":62,"rawValue":1391.336100037613,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2446,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=458","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2446,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1391.336100037613,"rating_lower":1377.9247004702859,"rating_upper":1404.7474996049402,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":46.8222221687138,"vote_count":2643},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1377.9247004702859,"type":"confidence_interval","upper":1404.7474996049402}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1391.336100037613,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c97f8ecc1fe98041a7bb","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-74a65e715a1b3dc0a1696824","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1404.864925632702,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2212,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=224","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2212,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1404.864925632702,"rating_lower":1392.6338193551348,"rating_upper":1417.096031910269,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.94352842395151,"vote_count":3108},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1392.6338193551348,"type":"confidence_interval","upper":1417.096031910269}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1404.864925632702,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0ad73576c6ee0e8e925e","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-69939919395603006e10798c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":63,"rawValue":1404.864925632702,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2050,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=62","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2050,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1404.864925632702,"rating_lower":1392.6338193551348,"rating_upper":1417.096031910269,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":38.94352842395151,"vote_count":3108},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1392.6338193551348,"type":"confidence_interval","upper":1417.096031910269}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1404.864925632702,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-819325ddced00be332b4","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7740fbf64ddd93a80f0ae0fd","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-instant","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":65,"rawValue":1414.20801751094,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2332,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=344","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2332,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1414.20801751094,"rating_lower":1385.4343682422448,"rating_upper":1442.9816667796354,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":215.52304238635327,"vote_count":465},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1385.4343682422448,"type":"confidence_interval","upper":1442.9816667796354}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1414.20801751094,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-874e04af73b4298e0bff","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f8c637d3ea6af02416523a67","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":34,"rawValue":1439.8067470314134,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2139,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=151","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2139,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1439.8067470314134,"rating_lower":1422.163937998912,"rating_upper":1457.4495560639148,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":81.02877710949228,"vote_count":1532},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1422.163937998912,"type":"confidence_interval","upper":1457.4495560639148}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1439.8067470314134,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c84f7283c3a0904bfcd7","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2ff1dcf6f17b22cb237591f8","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":51,"rawValue":1428.7607397003983,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2435,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=447","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2435,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1428.7607397003983,"rating_lower":1422.8522824147417,"rating_upper":1434.6691969860549,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.087658914464324,"vote_count":14859},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1422.8522824147417,"type":"confidence_interval","upper":1434.6691969860549}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1428.7607397003983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-46fbc6bf05447f77855c","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b17b5bb7638da6408e814066","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1435.9187347947436,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2041,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=53","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2041,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1435.9187347947436,"rating_lower":1430.5367451021693,"rating_upper":1441.300724487318,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.540315906794432,"vote_count":18367},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.5367451021693,"type":"confidence_interval","upper":1441.300724487318}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1435.9187347947436,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-25391d61be776316272f","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-61302701644af69d52161aad","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1435.9187347947436,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2203,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=215","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2203,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1435.9187347947436,"rating_lower":1430.5367451021693,"rating_upper":1441.300724487318,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.540315906794432,"vote_count":18367},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1430.5367451021693,"type":"confidence_interval","upper":1441.300724487318}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1435.9187347947436,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-67b729c6108e9bdf1aaf","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b8acf944f1941635bdbd48cd","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.5","modelRef":"kimi-k2.5-thinking","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":55,"rawValue":1429.8645129434199,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e3b9523b65a01c3094cd6f8c92b6249303822f0","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2322,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=334","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2322,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1429.8645129434199,"rating_lower":1416.9415661146954,"rating_upper":1442.787459772144,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.47373290542239,"vote_count":2376},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1416.9415661146954,"type":"confidence_interval","upper":1442.787459772144}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1429.8645129434199,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b87c5cd56e49a491ffd8","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5cb6e47a257e9989f912e6cf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":18,"rawValue":1520.953328879154,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb306229eb4434d3e0130c5169abca8942d0dfc8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2123,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=135","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2123,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1520.953328879154,"rating_lower":1509.347227396217,"rating_upper":1532.559430362091,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":35.06521816831401,"vote_count":3356},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1509.347227396217,"type":"confidence_interval","upper":1532.559430362091}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1520.953328879154,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-790f09bffe3aadbc1e94","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6665f09160c9b9bd91630dac","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1508.7146104959886,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb306229eb4434d3e0130c5169abca8942d0dfc8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2019,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=31","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2019,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1508.7146104959886,"rating_lower":1501.3003383528558,"rating_upper":1516.1288826391215,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":14.310040528431955,"vote_count":9589},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.3003383528558,"type":"confidence_interval","upper":1516.1288826391215}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1508.7146104959886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4e4eecff52e479e9c211","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4bb5130133e4f27c4d24f30e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1508.7146104959886,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb306229eb4434d3e0130c5169abca8942d0dfc8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2181,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=193","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2181,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1508.7146104959886,"rating_lower":1501.3003383528558,"rating_upper":1516.1288826391215,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":14.310040528431955,"vote_count":9589},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.3003383528558,"type":"confidence_interval","upper":1516.1288826391215}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1508.7146104959886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-761f9c9205aeaea34cb3","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-09e7ec3920c8bee8d0593e89","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":32,"rawValue":1503.6152719738125,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb306229eb4434d3e0130c5169abca8942d0dfc8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2416,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=428","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2416,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1503.6152719738125,"rating_lower":1495.4277700453038,"rating_upper":1511.8027739023214,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.45045071633066,"vote_count":6951},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1495.4277700453038,"type":"confidence_interval","upper":1511.8027739023214}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1503.6152719738125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cea3be64460c08c41f4b","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d79414915f9dd5731f4b7b02","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1494.9044004107018,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb306229eb4434d3e0130c5169abca8942d0dfc8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2304,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=316","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2304,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1494.9044004107018,"rating_lower":1476.0373035113412,"rating_upper":1513.7714973100628,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":92.66462612907843,"vote_count":1073},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1476.0373035113412,"type":"confidence_interval","upper":1513.7714973100628}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1494.9044004107018,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2e97339d86feb4aa01a4","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aee4bec6021ab9e9ac836256","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":20,"rawValue":1510.6217076220032,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d6fc4a24a614a1f12705b71c6981a7c5460d2ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2125,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=137","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2125,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1510.6217076220032,"rating_lower":1489.6254612008001,"rating_upper":1531.6179540432063,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":114.7591017779607,"vote_count":843},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1489.6254612008001,"type":"confidence_interval","upper":1531.6179540432063}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1510.6217076220032,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-842ceb8db7efa60c0940","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-30bfd548553b40d3ef082ef0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1472.3568398235943,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d6fc4a24a614a1f12705b71c6981a7c5460d2ab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2423,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=435","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2423,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1472.3568398235943,"rating_lower":1461.6265161661813,"rating_upper":1483.0871634810073,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.972948082268733,"vote_count":3761},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1461.6265161661813,"type":"confidence_interval","upper":1483.0871634810073}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1472.3568398235943,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-37f9343a9d35100b3f3e","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-326be36383b910e5b6a531d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1472.646928867663,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d6fc4a24a614a1f12705b71c6981a7c5460d2ab","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2028,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=40","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2028,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1472.646928867663,"rating_lower":1462.9592825115158,"rating_upper":1482.3345752238104,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.430950923174873,"vote_count":4680},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1462.9592825115158,"type":"confidence_interval","upper":1482.3345752238104}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1472.646928867663,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cde540d21712c60f9ff9","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-17bba06f3b9754f3d76de1b5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1472.646928867663,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d6fc4a24a614a1f12705b71c6981a7c5460d2ab","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2190,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=202","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2190,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1472.646928867663,"rating_lower":1462.9592825115158,"rating_upper":1482.3345752238104,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.430950923174873,"vote_count":4680},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1462.9592825115158,"type":"confidence_interval","upper":1482.3345752238104}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1472.646928867663,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-88530e869729680e16e3","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aad5a1f10ce3d93e9163426e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":49,"rawValue":1461.9566376410835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d6fc4a24a614a1f12705b71c6981a7c5460d2ab","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2316,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=328","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2316,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"moonshot","rating":1461.9566376410835,"rating_lower":1436.5187591137178,"rating_upper":1487.3945161684492,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":168.44789809723784,"vote_count":602},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1436.5187591137178,"type":"confidence_interval","upper":1487.3945161684492}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1461.9566376410835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-781397b325e8489dcc03","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f1e4ee64f40b1aab7db81f6c","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1673.6961333978227,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66fd4946324a1e8ccd81dfd33e7f56a80f2f8863","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1989,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=1","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1989,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1673.6961333978227,"rating_lower":1662.4419161479682,"rating_upper":1684.9503506476774,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.97117366574796,"vote_count":4546},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1662.4419161479682,"type":"confidence_interval","upper":1684.9503506476774}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1673.6961333978227,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-00a6f3eeddc6e16df60c","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-dbc37933b276967436c01ab6","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1673.6961333978227,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66fd4946324a1e8ccd81dfd33e7f56a80f2f8863","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2151,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=163","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2151,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1673.6961333978227,"rating_lower":1662.4419161479682,"rating_upper":1684.9503506476774,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":32.97117366574796,"vote_count":4546},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1662.4419161479682,"type":"confidence_interval","upper":1684.9503506476774}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1673.6961333978227,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1a72c685ba9d135dcab5","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-66be88ae65b1610eb702f484","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":2,"rawValue":1691.633912435193,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66fd4946324a1e8ccd81dfd33e7f56a80f2f8863","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2386,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=398","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2386,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1691.633912435193,"rating_lower":1677.8792829717245,"rating_upper":1705.388541898661,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":49.24947539672553,"vote_count":3278},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1677.8792829717245,"type":"confidence_interval","upper":1705.388541898661}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1691.633912435193,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-da2aa85c2fa7a2334658","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-00cfae1220900d9eeb130589","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1649.5392094087865,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66fd4946324a1e8ccd81dfd33e7f56a80f2f8863","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2270,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=282","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2270,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1649.5392094087865,"rating_lower":1616.8507924087075,"rating_upper":1682.2276264088655,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":278.158026897182,"vote_count":466},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1616.8507924087075,"type":"confidence_interval","upper":1682.2276264088655}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1649.5392094087865,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0372d01f933168319db8","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a5acf693c5c1ed7bc780cad2","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Kimi K3","modelRef":"kimi-k3-max","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","human_preference_rating","inferred_benchmark_version_hint"],"rank":8,"rawValue":1573.2857788791516,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66fd4946324a1e8ccd81dfd33e7f56a80f2f8863","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2113,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=125","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2113,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Kimi K3 license","observation_count":null,"organization":"moonshot","rating":1573.2857788791516,"rating_lower":1556.556981450059,"rating_upper":1590.0145763082446,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":72.85062172631841,"vote_count":1382},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1556.556981450059,"type":"confidence_interval","upper":1590.0145763082446}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1573.2857788791516,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ce265572f07a63869fbb","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9e0a99fc876bb121763922a7","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":88,"rawValue":1353.3448399219214,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2355,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=367","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2355,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1353.3448399219214,"rating_lower":1339.692541681641,"rating_upper":1366.9971381622022,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":48.51939222607429,"vote_count":3135},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1339.692541681641,"type":"confidence_interval","upper":1366.9971381622022}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1353.3448399219214,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d47a5be8a6efaf5f6ab0","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-c919707b3e2c6876e68fd8a5","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":89,"rawValue":1309.3574712037303,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2473,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=485","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2473,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1309.3574712037303,"rating_lower":1294.8486763736187,"rating_upper":1323.8662660338416,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":54.79822568662328,"vote_count":2277},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1294.8486763736187,"type":"confidence_interval","upper":1323.8662660338416}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1309.3574712037303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1621830e3755502f3410","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-bfe30c3935b7eb3b4209cda1","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":93,"rawValue":1330.1697248949574,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2080,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=92","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2080,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1330.1697248949574,"rating_lower":1320.2377673882402,"rating_upper":1340.1016824016747,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.678728972399472,"vote_count":5416},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1320.2377673882402,"type":"confidence_interval","upper":1340.1016824016747}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1330.1697248949574,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-107dc1c83964b38394f1","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-7d8da6bb50aad9ea3677fd6b","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (non-thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":93,"rawValue":1330.1697248949574,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2242,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=254","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2242,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1330.1697248949574,"rating_lower":1320.2377673882402,"rating_upper":1340.1016824016747,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":25.678728972399472,"vote_count":5416},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1320.2377673882402,"type":"confidence_interval","upper":1340.1016824016747}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1330.1697248949574,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-39f0bf80eddd56320c1d","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d1bf61fb5dfa39a541e8a0c0","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":92,"rawValue":1345.638787974788,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2359,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=371","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2359,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1345.638787974788,"rating_lower":1322.3099618621518,"rating_upper":1368.9676140874244,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":141.6738153906032,"vote_count":853},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1322.3099618621518,"type":"confidence_interval","upper":1368.9676140874244}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1345.638787974788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-144c056bce1c10238578","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-da66c4e9b35f4e4a792e5d88","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":97,"rawValue":1231.0262361783587,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2481,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=493","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2481,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1231.0262361783587,"rating_lower":1206.7934766971782,"rating_upper":1255.2589956595393,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":152.86552830120843,"vote_count":808},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1206.7934766971782,"type":"confidence_interval","upper":1255.2589956595393}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1231.0262361783587,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c97149dbae38965f9bed","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9556ebde6e61fba8112fd954","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":99,"rawValue":1292.4940956740304,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2248,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=260","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2248,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1292.4940956740304,"rating_lower":1275.9004307041535,"rating_upper":1309.0877606439074,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":71.67842478206467,"vote_count":1661},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1275.9004307041535,"type":"confidence_interval","upper":1309.0877606439074}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1292.4940956740304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9878e0636db3560aa4a2","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1fc261566dd053953c7c1de4","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash (thinking)","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":99,"rawValue":1292.4940956740304,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14be1dd4ec51e6746a690208cfd2db35e5568394","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2086,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=98","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2086,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1292.4940956740304,"rating_lower":1275.9004307041535,"rating_upper":1309.0877606439074,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":71.67842478206467,"vote_count":1661},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1275.9004307041535,"type":"confidence_interval","upper":1309.0877606439074}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1292.4940956740304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9d340e3e1ea75e7913ee","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b5e30a7ce03aaffaf50cfb8e","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1439.7163716626642,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80ea98de35fff6732e7d250f983022281e5c95e7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2320,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=332","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2320,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1439.7163716626642,"rating_lower":1418.2115083989104,"rating_upper":1461.221234926418,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":120.38633383272453,"vote_count":814},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1418.2115083989104,"type":"confidence_interval","upper":1461.221234926418}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1439.7163716626642,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-569b9363fbcdfb9b4975","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-572be7bb31b4dc9b3b018fa0","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1425.7960537879328,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80ea98de35fff6732e7d250f983022281e5c95e7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2437,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=449","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2437,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1425.7960537879328,"rating_lower":1416.9390519211117,"rating_upper":1434.6530556547536,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":20.42101340414607,"vote_count":6075},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1416.9390519211117,"type":"confidence_interval","upper":1434.6530556547536}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1425.7960537879328,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-870364b5b35e86f7f4c8","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-8e466987d2d9595ea59ad7d2","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1433.5094981427192,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80ea98de35fff6732e7d250f983022281e5c95e7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2044,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=56","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2044,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1433.5094981427192,"rating_lower":1425.2399418103105,"rating_upper":1441.7790544751276,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.801977094348356,"vote_count":6977},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1425.2399418103105,"type":"confidence_interval","upper":1441.7790544751276}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1433.5094981427192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-015ba3a55dec53f73497","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-50b434f18b8432d710694d8b","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":57,"rawValue":1433.5094981427192,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80ea98de35fff6732e7d250f983022281e5c95e7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2206,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=218","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2206,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"xiaomi","rating":1433.5094981427192,"rating_lower":1425.2399418103105,"rating_upper":1441.7790544751276,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.801977094348356,"vote_count":6977},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1425.2399418103105,"type":"confidence_interval","upper":1441.7790544751276}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1433.5094981427192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ea964359772056c9dc93","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6522d79c6cc527b968357fbc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1437.7972257239153,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92f0b7e460265a58ede9ea00733f2cd79bec7fe0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2202,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=214","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2202,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1437.7972257239153,"rating_lower":1431.3801087791285,"rating_upper":1444.2143426687023,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.719727011320531,"vote_count":12887},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1431.3801087791285,"type":"confidence_interval","upper":1444.2143426687023}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1437.7972257239153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9ef6dc83b00ef896fa20","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2c14a302838b0d0282be712c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":53,"rawValue":1437.7972257239153,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92f0b7e460265a58ede9ea00733f2cd79bec7fe0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2040,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=52","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2040,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1437.7972257239153,"rating_lower":1431.3801087791285,"rating_upper":1444.2143426687023,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.719727011320531,"vote_count":12887},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1431.3801087791285,"type":"confidence_interval","upper":1444.2143426687023}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1437.7972257239153,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-14062a903266fae62ee9","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-38b4491b6cd6cffa2898376c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":54,"rawValue":1425.36214121503,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92f0b7e460265a58ede9ea00733f2cd79bec7fe0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2438,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=450","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2438,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1425.36214121503,"rating_lower":1418.2850130026084,"rating_upper":1432.4392694274518,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":13.038209199391007,"vote_count":10071},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1418.2850130026084,"type":"confidence_interval","upper":1432.4392694274518}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1425.36214121503,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-4f5dad9b2df86b50f431","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-757b45f9b6e8cd04263e7dcf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":58,"rawValue":1428.1830694758214,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92f0b7e460265a58ede9ea00733f2cd79bec7fe0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2325,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=337","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2325,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1428.1830694758214,"rating_lower":1411.9869776516855,"rating_upper":1444.3791612999573,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":68.28483725056421,"vote_count":1500},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1411.9869776516855,"type":"confidence_interval","upper":1444.3791612999573}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1428.1830694758214,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-92cdb30bf6d6d5647a82","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f5b465e139ae27b1b142cea3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1475.6466788019832,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8472168ec1631067e4069d24656c426e7eb37ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2027,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=39","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2027,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1475.6466788019832,"rating_lower":1469.3799591036952,"rating_upper":1481.913398500271,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.223141158080029,"vote_count":13998},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1469.3799591036952,"type":"confidence_interval","upper":1481.913398500271}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1475.6466788019832,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f0f7616d74ef469adfa0","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cacaf63e1ac9ad4c5afba893","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1475.6466788019832,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8472168ec1631067e4069d24656c426e7eb37ca","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2189,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=201","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2189,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1475.6466788019832,"rating_lower":1469.3799591036952,"rating_upper":1481.913398500271,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.223141158080029,"vote_count":13998},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1469.3799591036952,"type":"confidence_interval","upper":1481.913398500271}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1475.6466788019832,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-7cf09cda75c0c3545216","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-496d6bc49586f727f5aaa4f8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1483.6255792907887,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8472168ec1631067e4069d24656c426e7eb37ca","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2307,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=319","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2307,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1483.6255792907887,"rating_lower":1467.8727966553379,"rating_upper":1499.378361926239,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":64.59789687786271,"vote_count":1602},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1467.8727966553379,"type":"confidence_interval","upper":1499.378361926239}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1483.6255792907887,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf67de4d359202769e26","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-035ce1e40fc92dac0bff4caa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":41,"rawValue":1464.0177570901815,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8472168ec1631067e4069d24656c426e7eb37ca","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2425,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=437","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2425,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"xiaomi","rating":1464.0177570901815,"rating_lower":1457.1512860826306,"rating_upper":1470.8842280977324,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.273572696795725,"vote_count":10938},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.1512860826306,"type":"confidence_interval","upper":1470.8842280977324}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1464.0177570901815,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a124debbe7adba860d07","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f564006ba6319898ac096e1f","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":69,"rawValue":1400.7095442722352,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88e06ae251ec03a34bae7233bf6ddc0d67513147","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2336,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=348","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2336,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1400.7095442722352,"rating_lower":1387.999802202853,"rating_upper":1413.4192863416172,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":42.05109335026812,"vote_count":5103},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1387.999802202853,"type":"confidence_interval","upper":1413.4192863416172}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1400.7095442722352,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6628785313578b208d98","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-28b503d2f0e3e059061c7048","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":72,"rawValue":1368.6020498541375,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88e06ae251ec03a34bae7233bf6ddc0d67513147","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2456,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=468","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2456,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1368.6020498541375,"rating_lower":1353.8299870394026,"rating_upper":1383.3741126688726,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":56.804940515552865,"vote_count":2186},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1353.8299870394026,"type":"confidence_interval","upper":1383.3741126688726}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1368.6020498541375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-d14319df6aaab0e7c684","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b86ad80a5ff8ddb54f00ecec","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":73,"rawValue":1386.9898745078774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88e06ae251ec03a34bae7233bf6ddc0d67513147","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2060,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=72","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2060,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1386.9898745078774,"rating_lower":1377.3180253350986,"rating_upper":1396.6617236806562,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.351339110301552,"vote_count":7296},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1377.3180253350986,"type":"confidence_interval","upper":1396.6617236806562}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1386.9898745078774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a627f857de988f1b67ad","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1d6767c731acbf23882f2c3c","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.1","modelRef":"minimax-m2.1-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":73,"rawValue":1386.9898745078774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88e06ae251ec03a34bae7233bf6ddc0d67513147","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2222,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=234","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2222,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MIT","observation_count":null,"organization":"minimax","rating":1386.9898745078774,"rating_lower":1377.3180253350986,"rating_upper":1396.6617236806562,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":24.351339110301552,"vote_count":7296},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1377.3180253350986,"type":"confidence_interval","upper":1396.6617236806562}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1386.9898745078774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-585964e6ebce966b1549","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-00fb96c01561c86605747f14","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":64,"rawValue":1414.5077245381826,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36f78a584a628e636de37d32fd438196ea2dc966","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2331,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=343","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2331,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1414.5077245381826,"rating_lower":1393.8341336291564,"rating_upper":1435.1813154472088,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":111.25912863398669,"vote_count":914},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1393.8341336291564,"type":"confidence_interval","upper":1435.1813154472088}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1414.5077245381826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-258815a99ef9bc746887","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1e3ad7029c58ddc900da6e41","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":71,"rawValue":1372.5005082678872,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36f78a584a628e636de37d32fd438196ea2dc966","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2455,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=467","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2455,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1372.5005082678872,"rating_lower":1363.4695777288882,"rating_upper":1381.5314388068862,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":21.230920389116033,"vote_count":6270},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1363.4695777288882,"type":"confidence_interval","upper":1381.5314388068862}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1372.5005082678872,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-83609d6fa886f0ff73bf","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f389b66021c04ba0d77ef978","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":75,"rawValue":1383.8983643309668,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36f78a584a628e636de37d32fd438196ea2dc966","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2224,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=236","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2224,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1383.8983643309668,"rating_lower":1375.4636645325552,"rating_upper":1392.3330641293787,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.520089374919326,"vote_count":7218},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1375.4636645325552,"type":"confidence_interval","upper":1392.3330641293787}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1383.8983643309668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-054862c8d34bd15b510b","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6696ec651ab7402f16badc1d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.5","modelRef":"minimax-m2.5","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":75,"rawValue":1383.8983643309668,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36f78a584a628e636de37d32fd438196ea2dc966","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2062,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=74","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2062,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1383.8983643309668,"rating_lower":1375.4636645325552,"rating_upper":1392.3330641293787,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.520089374919326,"vote_count":7218},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1375.4636645325552,"type":"confidence_interval","upper":1392.3330641293787}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1383.8983643309668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f870075af146fab90d33","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a50cf42bb1fd2a705644c343","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":64,"rawValue":1387.7484436814018,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcb9ac3d7c0821c218a0b9196c1bd3e9e4814ab9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2448,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=460","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2448,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1387.7484436814018,"rating_lower":1381.109988025986,"rating_upper":1394.3868993368176,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":11.47196821465812,"vote_count":11355},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1381.109988025986,"type":"confidence_interval","upper":1394.3868993368176}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1387.7484436814018,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a3ce9b8d27c276117941","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-68205c581788076d0aab9a60","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":67,"rawValue":1404.2259319703794,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcb9ac3d7c0821c218a0b9196c1bd3e9e4814ab9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2334,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=346","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2334,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1404.2259319703794,"rating_lower":1389.2168421975625,"rating_upper":1419.2350217431963,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":58.64250700669242,"vote_count":1718},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1389.2168421975625,"type":"confidence_interval","upper":1419.2350217431963}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1404.2259319703794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-96f2f5a88f4c35a1863c","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1f31eed35149f88e89a51aad","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":67,"rawValue":1398.13632454993,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcb9ac3d7c0821c218a0b9196c1bd3e9e4814ab9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2216,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=228","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2216,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1398.13632454993,"rating_lower":1392.0796137430818,"rating_upper":1404.1930353567777,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.549430961011845,"vote_count":14198},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1392.0796137430818,"type":"confidence_interval","upper":1404.1930353567777}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1398.13632454993,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a5deb3c61fe80990d2ad","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-099e15a5f903b24ac39e1570","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M2.7","modelRef":"minimax-m2.7","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":67,"rawValue":1398.13632454993,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcb9ac3d7c0821c218a0b9196c1bd3e9e4814ab9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2054,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=66","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2054,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Modified MIT","observation_count":null,"organization":"minimax","rating":1398.13632454993,"rating_lower":1392.0796137430818,"rating_upper":1404.1930353567777,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.549430961011845,"vote_count":14198},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1392.0796137430818,"type":"confidence_interval","upper":1404.1930353567777}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1398.13632454993,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-43f9c166b7de798dee7d","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-527644f2808f8d14a136bb34","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1472.4845412575187,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7c75bd1cb296f8794f03291be89b6cab81f474d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2132,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=144","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2132,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1472.4845412575187,"rating_lower":1460.9400936127213,"rating_upper":1484.028988902316,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":34.69366135216963,"vote_count":2956},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1460.9400936127213,"type":"confidence_interval","upper":1484.028988902316}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1472.4845412575187,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-72b9c8a310a523ea033f","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9755185a39fb818194d1bfb0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":35,"rawValue":1483.599854981654,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7c75bd1cb296f8794f03291be89b6cab81f474d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2419,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=431","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2419,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1483.599854981654,"rating_lower":1475.903578438817,"rating_upper":1491.2961315244913,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.419317344946375,"vote_count":8080},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1475.903578438817,"type":"confidence_interval","upper":1491.2961315244913}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1483.599854981654,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-82266fbb57d0f149236a","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6bb80820aa537a8defce02e4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1487.6424274606477,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7c75bd1cb296f8794f03291be89b6cab81f474d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2186,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=198","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2186,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1487.6424274606477,"rating_lower":1480.6862741015766,"rating_upper":1494.598580819719,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.596274439868736,"vote_count":10721},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1480.6862741015766,"type":"confidence_interval","upper":1494.598580819719}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1487.6424274606477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-396b4e31e31f745a9e4e","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-03eb8b537ee316cee8e4176e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":37,"rawValue":1487.6424274606477,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7c75bd1cb296f8794f03291be89b6cab81f474d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2024,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=36","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2024,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1487.6424274606477,"rating_lower":1480.6862741015766,"rating_upper":1494.598580819719,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":12.596274439868736,"vote_count":10721},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1480.6862741015766,"type":"confidence_interval","upper":1494.598580819719}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1487.6424274606477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-df09adbb1c1d3331e470","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aeac3023cb9538106c240026","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":46,"rawValue":1466.1016216771977,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7c75bd1cb296f8794f03291be89b6cab81f474d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2313,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=325","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2313,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"MiniMax Community License","observation_count":null,"organization":"minimax","rating":1466.1016216771977,"rating_lower":1448.3109446560525,"rating_upper":1483.892298698343,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":82.39270642852152,"vote_count":1282},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1448.3109446560525,"type":"confidence_interval","upper":1483.892298698343}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1466.1016216771977,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-823a8d02ab47ee636633","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3bd1acc0cb13686f7dc1ddfd","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":92,"rawValue":1256.494389060689,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"778b67b7e3ec9728d1b75b1ad2072d9ebee5ed66","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2476,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=488","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2476,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1256.494389060689,"rating_lower":1245.1807284554038,"rating_upper":1267.808049665974,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":33.32039265969567,"vote_count":3948},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1245.1807284554038,"type":"confidence_interval","upper":1267.808049665974}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1256.494389060689,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cad59496c66849cc692a","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-499cf2b6bb8bbc7ab2e91914","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":100,"rawValue":1273.2512988349522,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"778b67b7e3ec9728d1b75b1ad2072d9ebee5ed66","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2087,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=99","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2087,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1273.2512988349522,"rating_lower":1265.4290557422714,"rating_upper":1281.073541927633,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.928190267554731,"vote_count":12291},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1265.4290557422714,"type":"confidence_interval","upper":1281.073541927633}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1273.2512988349522,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e796067ab2563a65b4f1","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4592d2eb0ad3efe43bff2b14","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":100,"rawValue":1273.2512988349522,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"778b67b7e3ec9728d1b75b1ad2072d9ebee5ed66","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2249,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=261","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2249,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1273.2512988349522,"rating_lower":1265.4290557422714,"rating_upper":1281.073541927633,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.928190267554731,"vote_count":12291},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1265.4290557422714,"type":"confidence_interval","upper":1281.073541927633}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1273.2512988349522,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-07a1681103bb83fc1aee","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-562ac4c13c38e00720fe46f1","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Arena presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3-Coder 480B-A35B","modelRef":"qwen3-coder-480b-a35b-instruct","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":101,"rawValue":1286.321324993799,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"778b67b7e3ec9728d1b75b1ad2072d9ebee5ed66","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2368,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=380","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2368,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1286.321324993799,"rating_lower":1275.3109190806795,"rating_upper":1297.331730906918,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":31.558073125393005,"vote_count":8337},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1275.3109190806795,"type":"confidence_interval","upper":1297.331730906918}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1286.321324993799,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-aa7b9596d8b3e4bb8727","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-516bba21834f498cc65f7992","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":81,"rawValue":1349.5224759985815,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc9a7bcbce28bd489372abe451f298565defc772","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2465,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=477","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2465,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1349.5224759985815,"rating_lower":1341.2018620033107,"rating_upper":1357.8430899938521,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.02248064858469,"vote_count":6835},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1341.2018620033107,"type":"confidence_interval","upper":1357.8430899938521}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1349.5224759985815,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cfb05c01d242fb500f0f","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-aa9ebd87481100d7246c9eb2","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":84,"rawValue":1357.6812637830733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc9a7bcbce28bd489372abe451f298565defc772","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2233,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=245","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2233,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1357.6812637830733,"rating_lower":1349.8883753071802,"rating_upper":1365.4741522589663,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.808866795748525,"vote_count":7862},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1349.8883753071802,"type":"confidence_interval","upper":1365.4741522589663}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1357.6812637830733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-783a7afb2998b6efe349","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-3895b7f168df80886b9e2ed3","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":84,"rawValue":1357.6812637830733,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc9a7bcbce28bd489372abe451f298565defc772","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2071,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=83","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2071,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1357.6812637830733,"rating_lower":1349.8883753071802,"rating_upper":1365.4741522589663,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.808866795748525,"vote_count":7862},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1349.8883753071802,"type":"confidence_interval","upper":1365.4741522589663}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1357.6812637830733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-2a6952f821f956a1279f","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-0e58c4db37d04a732397e272","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 122B A10B","modelRef":"qwen3.5-122b-a10b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":87,"rawValue":1356.665847234471,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc9a7bcbce28bd489372abe451f298565defc772","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2354,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=366","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2354,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1356.665847234471,"rating_lower":1335.7618943146945,"rating_upper":1377.5698001542476,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":113.75242273019538,"vote_count":932},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1335.7618943146945,"type":"confidence_interval","upper":1377.5698001542476}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1356.665847234471,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8efe7b7909a5f41739a2","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-55964976da0651b501b9889c","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":74,"rawValue":1393.2231885877634,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7785999dbba990926f8ec322e2fb59bdec766a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2341,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=353","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2341,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1393.2231885877634,"rating_lower":1372.2379991117014,"rating_upper":1414.2083780638254,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":114.63826579993136,"vote_count":882},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1372.2379991117014,"type":"confidence_interval","upper":1414.2083780638254}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1393.2231885877634,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9bdba5b2459bb993bb09","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a4d21243ddcd583b02bf0750","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":83,"rawValue":1345.4015538321119,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7785999dbba990926f8ec322e2fb59bdec766a9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2467,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=479","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2467,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1345.4015538321119,"rating_lower":1336.7013593876775,"rating_upper":1354.1017482765462,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":19.70433288603851,"vote_count":6467},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1336.7013593876775,"type":"confidence_interval","upper":1354.1017482765462}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1345.4015538321119,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-73c44fef518307882fa5","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-a07f89c5807275152bb7eb05","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":85,"rawValue":1357.4623741792195,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7785999dbba990926f8ec322e2fb59bdec766a9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2072,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=84","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2072,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1357.4623741792195,"rating_lower":1349.2982234033402,"rating_upper":1365.6265249550988,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.351053597718316,"vote_count":7437},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1349.2982234033402,"type":"confidence_interval","upper":1365.6265249550988}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1357.4623741792195,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-201a4d2e60525b0af9af","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6157461d14b459de7375c628","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":85,"rawValue":1357.4623741792195,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7785999dbba990926f8ec322e2fb59bdec766a9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2234,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=246","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2234,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1357.4623741792195,"rating_lower":1349.2982234033402,"rating_upper":1365.6265249550988,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":17.351053597718316,"vote_count":7437},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1349.2982234033402,"type":"confidence_interval","upper":1365.6265249550988}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1357.4623741792195,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-fadeb2da6baad58adc92","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-99cca0d7443c1b29b5a6169e","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":95,"rawValue":1234.2886955279782,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0354a1be844619d350984d5855e7976c31c8cfc0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2479,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=491","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2479,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1234.2886955279782,"rating_lower":1215.083749038514,"rating_upper":1253.4936420174427,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":96.0129958119774,"vote_count":1352},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1215.083749038514,"type":"confidence_interval","upper":1253.4936420174427}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1234.2886955279782,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-02a2b49b702cfe1d96ce","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-44ddd6b24a71a662ebef7dba","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":99,"rawValue":1298.2173308353517,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0354a1be844619d350984d5855e7976c31c8cfc0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2366,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=378","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2366,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1298.2173308353517,"rating_lower":1252.7127975029935,"rating_upper":1343.72186416771,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":539.0302617955837,"vote_count":187},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1252.7127975029935,"type":"confidence_interval","upper":1343.72186416771}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1298.2173308353517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-a4b371cd93ac696b2bdd","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6414be94bd7f0e1d372e2e63","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":105,"rawValue":1249.749391429547,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0354a1be844619d350984d5855e7976c31c8cfc0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2092,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=104","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2092,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1249.749391429547,"rating_lower":1231.906788193733,"rating_upper":1267.5919946653605,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":82.87437275538667,"vote_count":1539},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1231.906788193733,"type":"confidence_interval","upper":1267.5919946653605}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1249.749391429547,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1c5eb3142bfad50b11aa","canonicalModelId":"qwen/qwen3.5-35b-a3b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5fa9ec92ebf66adfaf98cf4b","mappingCandidates":["qwen/qwen3.5-35b-a3b@current"],"mappingEvidence":{"confidence":"high","note":"Arena exact spelling of Qwen3.5 35B-A3B.","registry":"model_aliases.json"},"mappingNote":"Arena exact spelling of Qwen3.5 35B-A3B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 35B-A3B","modelRef":"qwen3.5-35b-a3b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":105,"rawValue":1249.749391429547,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0354a1be844619d350984d5855e7976c31c8cfc0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2254,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=266","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2254,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1249.749391429547,"rating_lower":1231.906788193733,"rating_upper":1267.5919946653605,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":82.87437275538667,"vote_count":1539},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1231.906788193733,"type":"confidence_interval","upper":1267.5919946653605}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1249.749391429547,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ea1787c01730b1b98f4c","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9380e5836c8fd136f7a62caa","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":65,"rawValue":1387.4312830150097,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3efaae8b7003302796914a0dd7a86bf2165082b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2449,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=461","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2449,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1387.4312830150097,"rating_lower":1381.4970767690395,"rating_upper":1393.3654892609798,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.167039245613378,"vote_count":14954},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1381.4970767690395,"type":"confidence_interval","upper":1393.3654892609798}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1387.4312830150097,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-03f3b47d9e0c158f661e","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f20f52f450efa767f3e869be","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":66,"rawValue":1399.2275095999478,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3efaae8b7003302796914a0dd7a86bf2165082b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2053,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=65","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2053,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1399.2275095999478,"rating_lower":1393.8269617234362,"rating_upper":1404.6280574764594,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.592406616303118,"vote_count":18612},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1393.8269617234362,"type":"confidence_interval","upper":1404.6280574764594}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1399.2275095999478,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f53aba5c0511df1deb73","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e11a1442445cda15a3c6ab4d","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":66,"rawValue":1399.2275095999478,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3efaae8b7003302796914a0dd7a86bf2165082b7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2215,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=227","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2215,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1399.2275095999478,"rating_lower":1393.8269617234362,"rating_upper":1404.6280574764594,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":7.592406616303118,"vote_count":18612},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1393.8269617234362,"type":"confidence_interval","upper":1404.6280574764594}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1399.2275095999478,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-f649b735071d478ce676","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-65de2e71f903c9e3f6fa45e7","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":66,"rawValue":1410.9917137840002,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3efaae8b7003302796914a0dd7a86bf2165082b7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2333,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=345","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2333,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1410.9917137840002,"rating_lower":1398.0566391336101,"rating_upper":1423.9267884343903,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":43.55536894208588,"vote_count":2245},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1398.0566391336101,"type":"confidence_interval","upper":1423.9267884343903}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1410.9917137840002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-cde5ce070b428a928ad3","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-64b23242f467c220f5068a51","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":98,"rawValue":1223.7875055486907,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd47adf0a98afbf599f1b1456b28a64c212d5f33","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2482,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=494","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2482,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1223.7875055486907,"rating_lower":1202.6107639540996,"rating_upper":1244.964247143282,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":116.74064606608393,"vote_count":1150},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1202.6107639540996,"type":"confidence_interval","upper":1244.964247143282}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1223.7875055486907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e22f687e1777eb1d593c","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-09ae4e132218620b57e3dcad","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":100,"rawValue":1293.103674188444,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd47adf0a98afbf599f1b1456b28a64c212d5f33","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2367,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=379","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2367,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1293.103674188444,"rating_lower":1232.6864510152218,"rating_upper":1353.5208973616664,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":950.2225655261306,"vote_count":121},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1232.6864510152218,"type":"confidence_interval","upper":1353.5208973616664}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1293.103674188444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-766f05e21f63e9d9c7b1","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9e8a446d7a8461edde754493","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":109,"rawValue":1238.1207909614184,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd47adf0a98afbf599f1b1456b28a64c212d5f33","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2096,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=108","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2096,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1238.1207909614184,"rating_lower":1218.0097541331572,"rating_upper":1258.2318277896798,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":105.2865125428061,"vote_count":1271},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1218.0097541331572,"type":"confidence_interval","upper":1258.2318277896798}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1238.1207909614184,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-61067ea6d51047e6df0b","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-70c393818856d06ba07dc393","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":109,"rawValue":1238.1207909614184,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd47adf0a98afbf599f1b1456b28a64c212d5f33","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2258,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=270","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2258,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1238.1207909614184,"rating_lower":1218.0097541331572,"rating_upper":1258.2318277896798,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":105.2865125428061,"vote_count":1271},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1218.0097541331572,"type":"confidence_interval","upper":1258.2318277896798}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1238.1207909614184,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b35adc490ef5f1ce1717","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-4edfd67f5ea943226aeefda7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":36,"rawValue":1497.0504325208572,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"710e300897c7ff60fcc34b6bb18fcd25920479f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2303,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=315","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2303,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1497.0504325208572,"rating_lower":1462.2498234669379,"rating_upper":1531.851041574776,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":315.2662691578373,"vote_count":323},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1462.2498234669379,"type":"confidence_interval","upper":1531.851041574776}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1497.0504325208572,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-01b1ef5ebaf6d8098dc0","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e898cb707a547d7aba20dedd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1478.8792606541383,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"710e300897c7ff60fcc34b6bb18fcd25920479f1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2188,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=200","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2188,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1478.8792606541383,"rating_lower":1466.1131319121796,"rating_upper":1491.6453893960972,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":42.42503972145128,"vote_count":2602},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1466.1131319121796,"type":"confidence_interval","upper":1491.6453893960972}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1478.8792606541383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-00bcb555abefadc9e057","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-18bb38cde9ff712109d202da","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":39,"rawValue":1478.8792606541383,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"710e300897c7ff60fcc34b6bb18fcd25920479f1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2026,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=38","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2026,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1478.8792606541383,"rating_lower":1466.1131319121796,"rating_upper":1491.6453893960972,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":42.42503972145128,"vote_count":2602},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1466.1131319121796,"type":"confidence_interval","upper":1491.6453893960972}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1478.8792606541383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-b1a8983dc8bd5194c818","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f50128c8cdccb93a5462411e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":40,"rawValue":1470.3582318681204,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"710e300897c7ff60fcc34b6bb18fcd25920479f1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2424,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=436","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2424,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1470.3582318681204,"rating_lower":1456.6008834050185,"rating_upper":1484.1155803312217,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":49.26894848270287,"vote_count":2237},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1456.6008834050185,"type":"confidence_interval","upper":1484.1155803312217}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1470.3582318681204,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-56108a5b63ef503d1102","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b7de192f4c99085ff437af48","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":28,"rawValue":1468.2702089284144,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dab13a7a3c50cab11a048031d945aa351acf56f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2133,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=145","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2133,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1468.2702089284144,"rating_lower":1457.7121329313518,"rating_upper":1478.8282849254774,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":29.018394824186316,"vote_count":5408},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1457.7121329313518,"type":"confidence_interval","upper":1478.8282849254774}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1468.2702089284144,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-bf05828b3e16d53b11a2","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9c02fbe78e8551c5343fc48d","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":44,"rawValue":1453.822661483878,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dab13a7a3c50cab11a048031d945aa351acf56f4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2428,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=440","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2428,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1453.822661483878,"rating_lower":1447.3266987470868,"rating_upper":1460.318624220669,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":10.984767466583056,"vote_count":12231},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.3266987470868,"type":"confidence_interval","upper":1460.318624220669}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1453.822661483878,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1c717864d6bfb1e5b3ac","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-d8aa1c3d17c6a470ca0d4be5","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1459.7403782936024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dab13a7a3c50cab11a048031d945aa351acf56f4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2194,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=206","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2194,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1459.7403782936024,"rating_lower":1453.8516329999622,"rating_upper":1465.6291235872425,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.0271229634327,"vote_count":16336},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1453.8516329999622,"type":"confidence_interval","upper":1465.6291235872425}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1459.7403782936024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-ea52236ddd1421d8132a","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-2fe89ebacedfa2f70d4136af","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":45,"rawValue":1459.7403782936024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dab13a7a3c50cab11a048031d945aa351acf56f4","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2032,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=44","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2032,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1459.7403782936024,"rating_lower":1453.8516329999622,"rating_upper":1465.6291235872425,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":9.0271229634327,"vote_count":16336},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1453.8516329999622,"type":"confidence_interval","upper":1465.6291235872425}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1459.7403782936024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-c8eea7148af85b4e13ee","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-1ae7a8d0afc0fb7595151d93","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":48,"rawValue":1462.49644395851,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dab13a7a3c50cab11a048031d945aa351acf56f4","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2315,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=327","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2315,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1462.49644395851,"rating_lower":1447.934129046901,"rating_upper":1477.0587588701187,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":55.20325102601149,"vote_count":1804},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1447.934129046901,"type":"confidence_interval","upper":1477.0587588701187}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1462.49644395851,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-aa4f630f42932ae62d82","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-15471b77846cc2aa0d6fc9f3","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-20260517","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":27,"rawValue":1514.03608266693,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8410afdfbe9e6117605411332216207386087f6a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2411,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=423","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2411,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1514.03608266693,"rating_lower":1505.5255489969084,"rating_upper":1522.5466163369515,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":18.854603610063787,"vote_count":6305},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1505.5255489969084,"type":"confidence_interval","upper":1522.5466163369515}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1514.03608266693,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-094a5d664af49f954ed5","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-f61536ea371f53563acc3f0c","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-20260517","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1516.6493621300187,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8410afdfbe9e6117605411332216207386087f6a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2017,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=29","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2017,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1516.6493621300187,"rating_lower":1509.054615097696,"rating_upper":1524.2441091623416,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.01517657152784,"vote_count":8380},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1509.054615097696,"type":"confidence_interval","upper":1524.2441091623416}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1516.6493621300187,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-6478b2fac0d04558bc1e","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-afc3735439dbc7e0c9985d3c","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-20260517","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":30,"rawValue":1516.6493621300187,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8410afdfbe9e6117605411332216207386087f6a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2179,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=191","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2179,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1516.6493621300187,"rating_lower":1509.054615097696,"rating_upper":1524.2441091623416,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":15.01517657152784,"vote_count":8380},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1509.054615097696,"type":"confidence_interval","upper":1524.2441091623416}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1516.6493621300187,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-900567363e9023370a8a","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-e931e4b11b60544554ef423f","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max-20260517","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","human_preference_rating","inferred_benchmark_version_hint"],"rank":31,"rawValue":1521.2481204137584,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8410afdfbe9e6117605411332216207386087f6a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2298,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=310","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2298,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1521.2481204137584,"rating_lower":1501.3354638130581,"rating_upper":1541.1607770144592,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":103.21961301820609,"vote_count":980},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1501.3354638130581,"type":"confidence_interval","upper":1541.1607770144592}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1521.2481204137584,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-217bd8a672c56198f7a5","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-ec5dc3a85da0ece7a3388ee7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":7,"rawValue":1573.5016936795942,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fffa35bf7589abab250ccf87266bc9a95783dfb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2112,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=124","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2112,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1573.5016936795942,"rating_lower":1558.3227179412515,"rating_upper":1588.6806694179368,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":59.97755415833512,"vote_count":1686},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1558.3227179412515,"type":"confidence_interval","upper":1588.6806694179368}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1573.5016936795942,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e5a9990b3d431d9d1302","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-cfe680009644a30ba26525f5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1579.4229632505258,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fffa35bf7589abab250ccf87266bc9a95783dfb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2276,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=288","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2276,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1579.4229632505258,"rating_lower":1541.1185351795934,"rating_upper":1617.7273913214576,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":381.9458383719175,"vote_count":267},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1541.1185351795934,"type":"confidence_interval","upper":1617.7273913214576}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1579.4229632505258,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-8f22746cd3bb1d468f2e","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b39eea8a1f2b9f23c0061e71","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1604.1093320820128,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fffa35bf7589abab250ccf87266bc9a95783dfb","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2393,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=405","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2393,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1604.1093320820128,"rating_lower":1587.6774284383198,"rating_upper":1620.5412357257057,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":70.28773962148612,"vote_count":1640},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1587.6774284383198,"type":"confidence_interval","upper":1620.5412357257057}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1604.1093320820128,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-0516046dad3c05363d26","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-633bd8e731b25e3c02716037","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1594.7775370883076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fffa35bf7589abab250ccf87266bc9a95783dfb","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1996,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=8","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1996,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1594.7775370883076,"rating_lower":1581.6995283838012,"rating_upper":1607.8555457928137,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.523270887031686,"vote_count":2464},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1581.6995283838012,"type":"confidence_interval","upper":1607.8555457928137}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1594.7775370883076,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-09cd7b4832cc9237f67c","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-5749757341d27bb136dd1bdd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":9,"rawValue":1594.7775370883076,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fffa35bf7589abab250ccf87266bc9a95783dfb","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2158,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=170","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2158,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Apache 2.0","observation_count":null,"organization":"alibaba","rating":1594.7775370883076,"rating_lower":1581.6995283838012,"rating_upper":1607.8555457928137,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":44.523270887031686,"vote_count":2464},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1581.6995283838012,"type":"confidence_interval","upper":1607.8555457928137}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1594.7775370883076,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-1df4f5f18bf59b07700d","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-fecbf82fc2d37fe1b508116f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1668.9390977471774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b80e7e17168d5a6a8e0058895537e1971c8c0fc8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2152,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=164","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2152,"metadata":{"category":"webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1668.9390977471774,"rating_lower":1656.4367273997766,"rating_upper":1681.441468094578,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.69007936816081,"vote_count":3212},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1656.4367273997766,"type":"confidence_interval","upper":1681.441468094578}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1668.9390977471774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-e1ff0cf52a994aa5a7d9","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-b9033f9b36c7977546b02681","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"overall","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1668.9390977471774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b80e7e17168d5a6a8e0058895537e1971c8c0fc8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1990,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=2","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":1990,"metadata":{"category":"overall","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1668.9390977471774,"rating_lower":1656.4367273997766,"rating_upper":1681.441468094578,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":40.69007936816081,"vote_count":3212},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1656.4367273997766,"type":"confidence_interval","upper":1681.441468094578}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1668.9390977471774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-5491411e0c4edf2aa318","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-9066c8932e8ce5f2f830383a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-25","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"image_to_webdev","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":3,"rawValue":1618.4401641653608,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b80e7e17168d5a6a8e0058895537e1971c8c0fc8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2108,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=120","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2108,"metadata":{"category":"image_to_webdev","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1618.4401641653608,"rating_lower":1603.5329009090706,"rating_upper":1633.347427421651,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.84950670177612,"vote_count":1869},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1603.5329009090706,"type":"confidence_interval","upper":1633.347427421651}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1618.4401641653608,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-06a913a80a73639e83dc","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-56ed1e99e1c853d6bf51c1ea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-react","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":4,"rawValue":1678.4351937220601,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b80e7e17168d5a6a8e0058895537e1971c8c0fc8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2388,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=400","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2388,"metadata":{"category":"webdev-react","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1678.4351937220601,"rating_lower":1663.6333658140602,"rating_upper":1693.23702163006,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":57.03409034030969,"vote_count":2322},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1663.6333658140602,"type":"confidence_interval","upper":1693.23702163006}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1678.4351937220601,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"arena-webdev","benchmarkName":"Arena WebDev","benchmarkVersion":null,"benchmarkVersionHint":"arena-webdev@latest","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-lmarena-hf-dataset-9e8138bd923bc71b7147","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://datasets-server.huggingface.co/rows","harness":"arena-human-preference","harnessId":"arena-human-preference","httpStatus":200,"id":"pub-6e59474f534280e5fbbcfa26","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"arena_score_bt","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026-08-21","parserVersion":"0.1.0","payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","protocol":{"arena_config":"webdev","category":"webdev-html","harness":"arena-human-preference","rating_method":"bradley-terry","split":"latest","subject_type":"system"},"publishedAt":null,"qualityFlags":["human_preference_rating","inferred_benchmark_version_hint"],"rank":6,"rawValue":1627.719515798759,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b80e7e17168d5a6a8e0058895537e1971c8c0fc8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2273,"payloadSha256":"b4216be476f616e31837b47784db8c4591628bec33ffdfb2cc2f3ea96b924557","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://datasets-server.huggingface.co/rows","sourceId":"lmarena-hf-dataset","sourceLabel":"Arena · official Hugging Face leaderboard dataset","sourceLocator":"dataset=lmarena-ai/leaderboard-dataset;config=webdev;split=latest;row_idx=285","sourceMetricId":"elo","sourcePageUrl":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","sourceRow":{"artifact":"lmarena-hf-dataset/candidates.jsonl","line":2273,"metadata":{"category":"webdev-html","config":"webdev","dataset":"lmarena-ai/leaderboard-dataset","license":"Proprietary","observation_count":null,"organization":"alibaba","rating":1627.719515798759,"rating_lower":1593.125429491894,"rating_upper":1662.3136021056239,"score":null,"score_ci_lower":null,"score_ci_upper":null,"session_count":null,"source_status":"publisher_dataset","split":"latest","variance":311.5355033769619,"vote_count":396},"sourceStatus":"publisher_dataset","uncertainty":{"level":0.95,"lower":1593.125429491894,"type":"confidence_interval","upper":1662.3136021056239}},"sourceSubjectType":"system","sourceUrl":"https://datasets-server.huggingface.co/rows","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"rating","value":1627.719515798759,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-d82250ef199aa77eecb0","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-7f2979335561d82026ffa97e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Claude Haiku 4.5","modelRef":"Claude-Haiku-4-5-20251001","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":6,"rawValue":"68.70%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c1bbad98e2be51b4e6da038614a7943e02b6bd3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":6,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=5;rank=6;model=Claude-Haiku-4-5-20251001 (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":6,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"1.68","latency_mean_s":1.68,"latency_p95_raw":"3.15","latency_p95_s":3.15,"latency_std_raw":"3.92","latency_std_s":3.92,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Claude-Haiku-4-5-20251001","model_link":"https://www.anthropic.com/news/claude-haiku-4-5","organization":"Anthropic","overall_accuracy":"68.70%","overall_accuracy_percent":68.7,"rank":6,"source_model":"Claude-Haiku-4-5-20251001 (FC)","source_row_index":5,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"85.11%","value":85.11},"Live Acc":{"raw":"78.68%","value":78.68},"Live Multiple AST":{"raw":"77.59%","value":77.59},"Live Parallel AST":{"raw":"75.00%","value":75.0},"Live Parallel Multiple AST":{"raw":"75.00%","value":75.0},"Live Simple AST":{"raw":"83.72%","value":83.72},"Memory Acc":{"raw":"54.41%","value":54.41},"Memory KV":{"raw":"51.61%","value":51.61},"Memory Recursive Summarization":{"raw":"56.13%","value":56.13},"Memory Vector":{"raw":"55.48%","value":55.48},"Multi Turn Acc":{"raw":"53.62%","value":53.62},"Multi Turn Base":{"raw":"63.50%","value":63.5},"Multi Turn Long Context":{"raw":"56.00%","value":56.0},"Multi Turn Miss Func":{"raw":"42.50%","value":42.5},"Multi Turn Miss Param":{"raw":"52.50%","value":52.5},"Non-Live AST Acc":{"raw":"86.50%","value":86.5},"Non-Live Multiple AST":{"raw":"94.00%","value":94.0},"Non-Live Parallel AST":{"raw":"92.50%","value":92.5},"Non-Live Parallel Multiple AST":{"raw":"88.50%","value":88.5},"Non-Live Simple AST":{"raw":"71.00%","value":71.0},"Relevance Detection":{"raw":"62.50%","value":62.5},"Web Search Acc":{"raw":"83.50%","value":83.5},"Web Search Base":{"raw":"86.00%","value":86.0},"Web Search No Snippet":{"raw":"81.00%","value":81.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"85.11%","Live Acc":"78.68%","Live Multiple AST":"77.59%","Live Parallel AST":"75.00%","Live Parallel Multiple AST":"75.00%","Live Simple AST":"83.72%","Memory Acc":"54.41%","Memory KV":"51.61%","Memory Recursive Summarization":"56.13%","Memory Vector":"55.48%","Multi Turn Acc":"53.62%","Multi Turn Base":"63.50%","Multi Turn Long Context":"56.00%","Multi Turn Miss Func":"42.50%","Multi Turn Miss Param":"52.50%","Non-Live AST Acc":"86.50%","Non-Live Multiple AST":"94.00%","Non-Live Parallel AST":"92.50%","Non-Live Parallel Multiple AST":"88.50%","Non-Live Simple AST":"71.00%","Relevance Detection":"62.50%","Web Search Acc":"83.50%","Web Search Base":"86.00%","Web Search No Snippet":"81.00%"},"total_cost_raw":"14.23","total_cost_usd":14.23,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":68.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-4ab86139f0ff3b068d88","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-226e95ae5dc15e984a9bae2f","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Claude Haiku 4.5","modelRef":"Claude-Haiku-4-5-20251001","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":87,"rawValue":"25.26%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c1bbad98e2be51b4e6da038614a7943e02b6bd3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":87,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=86;rank=87;model=Claude-Haiku-4-5-20251001 (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":87,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"3.75","latency_mean_s":3.75,"latency_p95_raw":"3.77","latency_p95_s":3.77,"latency_std_raw":"19.96","latency_std_s":19.96,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Claude-Haiku-4-5-20251001","model_link":"https://www.anthropic.com/news/claude-haiku-4-5","organization":"Anthropic","overall_accuracy":"25.26%","overall_accuracy_percent":25.26,"rank":87,"source_model":"Claude-Haiku-4-5-20251001 (Prompt)","source_row_index":86,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"67.5","value":67.5},"Format Sensitivity Standard Deviation":{"raw":"20.07","value":20.07},"Irrelevance Detection":{"raw":"95.29%","value":95.29},"Live Acc":{"raw":"52.48%","value":52.48},"Live Multiple AST":{"raw":"49.76%","value":49.76},"Live Parallel AST":{"raw":"56.25%","value":56.25},"Live Parallel Multiple AST":{"raw":"16.67%","value":16.67},"Live Simple AST":{"raw":"66.67%","value":66.67},"Memory Acc":{"raw":"2.58%","value":2.58},"Memory KV":{"raw":"2.58%","value":2.58},"Memory Recursive Summarization":{"raw":"3.23%","value":3.23},"Memory Vector":{"raw":"1.94%","value":1.94},"Multi Turn Acc":{"raw":"1.75%","value":1.75},"Multi Turn Base":{"raw":"1.50%","value":1.5},"Multi Turn Long Context":{"raw":"1.50%","value":1.5},"Multi Turn Miss Func":{"raw":"0.00%","value":0.0},"Multi Turn Miss Param":{"raw":"4.00%","value":4.0},"Non-Live AST Acc":{"raw":"55.42%","value":55.42},"Non-Live Multiple AST":{"raw":"84.00%","value":84.0},"Non-Live Parallel AST":{"raw":"38.00%","value":38.0},"Non-Live Parallel Multiple AST":{"raw":"44.00%","value":44.0},"Non-Live Simple AST":{"raw":"55.67%","value":55.67},"Relevance Detection":{"raw":"31.25%","value":31.25},"Web Search Acc":{"raw":"19.50%","value":19.5},"Web Search Base":{"raw":"20.00%","value":20.0},"Web Search No Snippet":{"raw":"19.00%","value":19.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"67.5","Format Sensitivity Standard Deviation":"20.07","Irrelevance Detection":"95.29%","Live Acc":"52.48%","Live Multiple AST":"49.76%","Live Parallel AST":"56.25%","Live Parallel Multiple AST":"16.67%","Live Simple AST":"66.67%","Memory Acc":"2.58%","Memory KV":"2.58%","Memory Recursive Summarization":"3.23%","Memory Vector":"1.94%","Multi Turn Acc":"1.75%","Multi Turn Base":"1.50%","Multi Turn Long Context":"1.50%","Multi Turn Miss Func":"0.00%","Multi Turn Miss Param":"4.00%","Non-Live AST Acc":"55.42%","Non-Live Multiple AST":"84.00%","Non-Live Parallel AST":"38.00%","Non-Live Parallel Multiple AST":"44.00%","Non-Live Simple AST":"55.67%","Relevance Detection":"31.25%","Web Search Acc":"19.50%","Web Search Base":"20.00%","Web Search No Snippet":"19.00%"},"total_cost_raw":"45.13","total_cost_usd":45.13,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":25.26,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-b129e8da985086928679","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-2226383995d7f954e72fa4f2","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Claude Opus 4.5","modelRef":"Claude-Opus-4-5-20251101","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":1,"rawValue":"77.47%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cee834c0003b6639776fb5a4974f6e64d29f173e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":1,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=0;rank=1;model=Claude-Opus-4-5-20251101 (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":1,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"4.38","latency_mean_s":4.38,"latency_p95_raw":"7.56","latency_p95_s":7.56,"latency_std_raw":"3.13","latency_std_s":3.13,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Claude-Opus-4-5-20251101","model_link":"https://www.anthropic.com/news/claude-4","organization":"Anthropic","overall_accuracy":"77.47%","overall_accuracy_percent":77.47,"rank":1,"source_model":"Claude-Opus-4-5-20251101 (FC)","source_row_index":0,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"84.72%","value":84.72},"Live Acc":{"raw":"79.79%","value":79.79},"Live Multiple AST":{"raw":"78.16%","value":78.16},"Live Parallel AST":{"raw":"87.50%","value":87.5},"Live Parallel Multiple AST":{"raw":"75.00%","value":75.0},"Live Simple AST":{"raw":"86.43%","value":86.43},"Memory Acc":{"raw":"73.76%","value":73.76},"Memory KV":{"raw":"70.97%","value":70.97},"Memory Recursive Summarization":{"raw":"77.42%","value":77.42},"Memory Vector":{"raw":"72.90%","value":72.9},"Multi Turn Acc":{"raw":"68.38%","value":68.38},"Multi Turn Base":{"raw":"81.00%","value":81.0},"Multi Turn Long Context":{"raw":"70.50%","value":70.5},"Multi Turn Miss Func":{"raw":"64.00%","value":64.0},"Multi Turn Miss Param":{"raw":"58.00%","value":58.0},"Non-Live AST Acc":{"raw":"88.58%","value":88.58},"Non-Live Multiple AST":{"raw":"95.50%","value":95.5},"Non-Live Parallel AST":{"raw":"93.50%","value":93.5},"Non-Live Parallel Multiple AST":{"raw":"88.50%","value":88.5},"Non-Live Simple AST":{"raw":"76.83%","value":76.83},"Relevance Detection":{"raw":"62.50%","value":62.5},"Web Search Acc":{"raw":"84.50%","value":84.5},"Web Search Base":{"raw":"84.00%","value":84.0},"Web Search No Snippet":{"raw":"85.00%","value":85.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"84.72%","Live Acc":"79.79%","Live Multiple AST":"78.16%","Live Parallel AST":"87.50%","Live Parallel Multiple AST":"75.00%","Live Simple AST":"86.43%","Memory Acc":"73.76%","Memory KV":"70.97%","Memory Recursive Summarization":"77.42%","Memory Vector":"72.90%","Multi Turn Acc":"68.38%","Multi Turn Base":"81.00%","Multi Turn Long Context":"70.50%","Multi Turn Miss Func":"64.00%","Multi Turn Miss Param":"58.00%","Non-Live AST Acc":"88.58%","Non-Live Multiple AST":"95.50%","Non-Live Parallel AST":"93.50%","Non-Live Parallel Multiple AST":"88.50%","Non-Live Simple AST":"76.83%","Relevance Detection":"62.50%","Web Search Acc":"84.50%","Web Search Base":"84.00%","Web Search No Snippet":"85.00%"},"total_cost_raw":"86.55","total_cost_usd":86.55,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":77.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-66a5f7f88593f51f8b96","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-2f6a6383db8cf0165ac4a8d9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Claude Opus 4.5","modelRef":"Claude-Opus-4-5-20251101","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":57,"rawValue":"33.47%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cee834c0003b6639776fb5a4974f6e64d29f173e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":57,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=56;rank=57;model=Claude-Opus-4-5-20251101 (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":57,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"3.76","latency_mean_s":3.76,"latency_p95_raw":"5.52","latency_p95_s":5.52,"latency_std_raw":"13.19","latency_std_s":13.19,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Claude-Opus-4-5-20251101","model_link":"https://www.anthropic.com/news/claude-4","organization":"Anthropic","overall_accuracy":"33.47%","overall_accuracy_percent":33.47,"rank":57,"source_model":"Claude-Opus-4-5-20251101 (Prompt)","source_row_index":56,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"13.0","value":13.0},"Format Sensitivity Standard Deviation":{"raw":"3.65","value":3.65},"Irrelevance Detection":{"raw":"90.75%","value":90.75},"Live Acc":{"raw":"76.02%","value":76.02},"Live Multiple AST":{"raw":"74.17%","value":74.17},"Live Parallel AST":{"raw":"81.25%","value":81.25},"Live Parallel Multiple AST":{"raw":"62.50%","value":62.5},"Live Simple AST":{"raw":"84.50%","value":84.5},"Memory Acc":{"raw":"1.94%","value":1.94},"Memory KV":{"raw":"1.29%","value":1.29},"Memory Recursive Summarization":{"raw":"2.58%","value":2.58},"Memory Vector":{"raw":"1.94%","value":1.94},"Multi Turn Acc":{"raw":"16.12%","value":16.12},"Multi Turn Base":{"raw":"20.50%","value":20.5},"Multi Turn Long Context":{"raw":"13.50%","value":13.5},"Multi Turn Miss Func":{"raw":"9.00%","value":9.0},"Multi Turn Miss Param":{"raw":"21.50%","value":21.5},"Non-Live AST Acc":{"raw":"89.65%","value":89.65},"Non-Live Multiple AST":{"raw":"93.50%","value":93.5},"Non-Live Parallel AST":{"raw":"93.00%","value":93.0},"Non-Live Parallel Multiple AST":{"raw":"92.50%","value":92.5},"Non-Live Simple AST":{"raw":"79.58%","value":79.58},"Relevance Detection":{"raw":"68.75%","value":68.75},"Web Search Acc":{"raw":"13.00%","value":13.0},"Web Search Base":{"raw":"13.00%","value":13.0},"Web Search No Snippet":{"raw":"13.00%","value":13.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"13.0","Format Sensitivity Standard Deviation":"3.65","Irrelevance Detection":"90.75%","Live Acc":"76.02%","Live Multiple AST":"74.17%","Live Parallel AST":"81.25%","Live Parallel Multiple AST":"62.50%","Live Simple AST":"84.50%","Memory Acc":"1.94%","Memory KV":"1.29%","Memory Recursive Summarization":"2.58%","Memory Vector":"1.94%","Multi Turn Acc":"16.12%","Multi Turn Base":"20.50%","Multi Turn Long Context":"13.50%","Multi Turn Miss Func":"9.00%","Multi Turn Miss Param":"21.50%","Non-Live AST Acc":"89.65%","Non-Live Multiple AST":"93.50%","Non-Live Parallel AST":"93.00%","Non-Live Parallel Multiple AST":"92.50%","Non-Live Simple AST":"79.58%","Relevance Detection":"68.75%","Web Search Acc":"13.00%","Web Search Base":"13.00%","Web Search No Snippet":"13.00%"},"total_cost_raw":"88.33","total_cost_usd":88.33,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":33.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-fdaf6a39ebdf7a379b71","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-91209b75e3bc183243a45082","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Claude Sonnet 4.5","modelRef":"Claude-Sonnet-4-5-20250929","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":2,"rawValue":"73.24%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a49fede302be97368593f87e769c51a512124d82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":2,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=1;rank=2;model=Claude-Sonnet-4-5-20250929 (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":2,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"4.31","latency_mean_s":4.31,"latency_p95_raw":"7.27","latency_p95_s":7.27,"latency_std_raw":"4.43","latency_std_s":4.43,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Claude-Sonnet-4-5-20250929","model_link":"https://www.anthropic.com/news/claude-sonnet-4-5","organization":"Anthropic","overall_accuracy":"73.24%","overall_accuracy_percent":73.24,"rank":2,"source_model":"Claude-Sonnet-4-5-20250929 (FC)","source_row_index":1,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"86.61%","value":86.61},"Live Acc":{"raw":"81.13%","value":81.13},"Live Multiple AST":{"raw":"78.92%","value":78.92},"Live Parallel AST":{"raw":"87.50%","value":87.5},"Live Parallel Multiple AST":{"raw":"83.33%","value":83.33},"Live Simple AST":{"raw":"89.53%","value":89.53},"Memory Acc":{"raw":"64.95%","value":64.95},"Memory KV":{"raw":"54.19%","value":54.19},"Memory Recursive Summarization":{"raw":"83.23%","value":83.23},"Memory Vector":{"raw":"57.42%","value":57.42},"Multi Turn Acc":{"raw":"61.37%","value":61.37},"Multi Turn Base":{"raw":"69.00%","value":69.0},"Multi Turn Long Context":{"raw":"59.00%","value":59.0},"Multi Turn Miss Func":{"raw":"65.00%","value":65.0},"Multi Turn Miss Param":{"raw":"52.50%","value":52.5},"Non-Live AST Acc":{"raw":"88.65%","value":88.65},"Non-Live Multiple AST":{"raw":"95.50%","value":95.5},"Non-Live Parallel AST":{"raw":"94.50%","value":94.5},"Non-Live Parallel Multiple AST":{"raw":"92.00%","value":92.0},"Non-Live Simple AST":{"raw":"72.58%","value":72.58},"Relevance Detection":{"raw":"68.75%","value":68.75},"Web Search Acc":{"raw":"81.00%","value":81.0},"Web Search Base":{"raw":"82.00%","value":82.0},"Web Search No Snippet":{"raw":"80.00%","value":80.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"86.61%","Live Acc":"81.13%","Live Multiple AST":"78.92%","Live Parallel AST":"87.50%","Live Parallel Multiple AST":"83.33%","Live Simple AST":"89.53%","Memory Acc":"64.95%","Memory KV":"54.19%","Memory Recursive Summarization":"83.23%","Memory Vector":"57.42%","Multi Turn Acc":"61.37%","Multi Turn Base":"69.00%","Multi Turn Long Context":"59.00%","Multi Turn Miss Func":"65.00%","Multi Turn Miss Param":"52.50%","Non-Live AST Acc":"88.65%","Non-Live Multiple AST":"95.50%","Non-Live Parallel AST":"94.50%","Non-Live Parallel Multiple AST":"92.00%","Non-Live Simple AST":"72.58%","Relevance Detection":"68.75%","Web Search Acc":"81.00%","Web Search Base":"82.00%","Web Search No Snippet":"80.00%"},"total_cost_raw":"43.73","total_cost_usd":43.73,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":73.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-25844d89cab25afcf1c0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-bc605ce72442684036fcba27","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Claude Sonnet 4.5","modelRef":"Claude-Sonnet-4-5-20250929","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":89,"rawValue":"24.90%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a49fede302be97368593f87e769c51a512124d82","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":89,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=88;rank=89;model=Claude-Sonnet-4-5-20250929 (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":89,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"3.84","latency_mean_s":3.84,"latency_p95_raw":"6.66","latency_p95_s":6.66,"latency_std_raw":"1.53","latency_std_s":1.53,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Claude-Sonnet-4-5-20250929","model_link":"https://www.anthropic.com/news/claude-sonnet-4-5","organization":"Anthropic","overall_accuracy":"24.90%","overall_accuracy_percent":24.9,"rank":89,"source_model":"Claude-Sonnet-4-5-20250929 (Prompt)","source_row_index":88,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"37.5","value":37.5},"Format Sensitivity Standard Deviation":{"raw":"10.07","value":10.07},"Irrelevance Detection":{"raw":"95.03%","value":95.03},"Live Acc":{"raw":"46.56%","value":46.56},"Live Multiple AST":{"raw":"40.17%","value":40.17},"Live Parallel AST":{"raw":"56.25%","value":56.25},"Live Parallel Multiple AST":{"raw":"33.33%","value":33.33},"Live Simple AST":{"raw":"73.26%","value":73.26},"Memory Acc":{"raw":"5.38%","value":5.38},"Memory KV":{"raw":"4.52%","value":4.52},"Memory Recursive Summarization":{"raw":"1.94%","value":1.94},"Memory Vector":{"raw":"9.68%","value":9.68},"Multi Turn Acc":{"raw":"1.62%","value":1.62},"Multi Turn Base":{"raw":"2.00%","value":2.0},"Multi Turn Long Context":{"raw":"1.50%","value":1.5},"Multi Turn Miss Func":{"raw":"0.00%","value":0.0},"Multi Turn Miss Param":{"raw":"3.00%","value":3.0},"Non-Live AST Acc":{"raw":"59.81%","value":59.81},"Non-Live Multiple AST":{"raw":"79.50%","value":79.5},"Non-Live Parallel AST":{"raw":"53.50%","value":53.5},"Non-Live Parallel Multiple AST":{"raw":"59.00%","value":59.0},"Non-Live Simple AST":{"raw":"47.25%","value":47.25},"Relevance Detection":{"raw":"37.50%","value":37.5},"Web Search Acc":{"raw":"16.00%","value":16.0},"Web Search Base":{"raw":"16.00%","value":16.0},"Web Search No Snippet":{"raw":"16.00%","value":16.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"37.5","Format Sensitivity Standard Deviation":"10.07","Irrelevance Detection":"95.03%","Live Acc":"46.56%","Live Multiple AST":"40.17%","Live Parallel AST":"56.25%","Live Parallel Multiple AST":"33.33%","Live Simple AST":"73.26%","Memory Acc":"5.38%","Memory KV":"4.52%","Memory Recursive Summarization":"1.94%","Memory Vector":"9.68%","Multi Turn Acc":"1.62%","Multi Turn Base":"2.00%","Multi Turn Long Context":"1.50%","Multi Turn Miss Func":"0.00%","Multi Turn Miss Param":"3.00%","Non-Live AST Acc":"59.81%","Non-Live Multiple AST":"79.50%","Non-Live Parallel AST":"53.50%","Non-Live Parallel Multiple AST":"59.00%","Non-Live Simple AST":"47.25%","Relevance Detection":"37.50%","Web Search Acc":"16.00%","Web Search Base":"16.00%","Web Search No Snippet":"16.00%"},"total_cost_raw":"47.82","total_cost_usd":47.82,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":24.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-08f9f4da238caf2b434e","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-26ce2bdb7c58284b460e27cf","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt + Thinking","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":"Thinking"},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":14,"rawValue":"56.73%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"671ad7005b9522d9d14de8d1613e3e515ee79bf0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":14,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=13;rank=14;model=DeepSeek-V3.2-Exp (Prompt + Thinking);mode=Prompt + Thinking","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":14,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt + Thinking","evaluator_commit":"f7cf735","latency_mean_raw":"37.89","latency_mean_s":37.89,"latency_p95_raw":"102.09","latency_p95_s":102.09,"latency_std_raw":"49.56","latency_std_s":49.56,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"MIT","model":"DeepSeek-V3.2-Exp","model_link":"https://api-docs.deepseek.com/news/news250528","organization":"DeepSeek","overall_accuracy":"56.73%","overall_accuracy_percent":56.73,"rank":14,"source_model":"DeepSeek-V3.2-Exp (Prompt + Thinking)","source_row_index":13,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"10.0","value":10.0},"Format Sensitivity Standard Deviation":{"raw":"2.77","value":2.77},"Irrelevance Detection":{"raw":"67.00%","value":67.0},"Live Acc":{"raw":"76.02%","value":76.02},"Live Multiple AST":{"raw":"74.74%","value":74.74},"Live Parallel AST":{"raw":"87.50%","value":87.5},"Live Parallel Multiple AST":{"raw":"54.17%","value":54.17},"Live Simple AST":{"raw":"82.56%","value":82.56},"Memory Acc":{"raw":"44.09%","value":44.09},"Memory KV":{"raw":"46.45%","value":46.45},"Memory Recursive Summarization":{"raw":"39.35%","value":39.35},"Memory Vector":{"raw":"46.45%","value":46.45},"Multi Turn Acc":{"raw":"44.88%","value":44.88},"Multi Turn Base":{"raw":"55.00%","value":55.0},"Multi Turn Long Context":{"raw":"48.50%","value":48.5},"Multi Turn Miss Func":{"raw":"49.00%","value":49.0},"Multi Turn Miss Param":{"raw":"27.00%","value":27.0},"Non-Live AST Acc":{"raw":"85.52%","value":85.52},"Non-Live Multiple AST":{"raw":"92.00%","value":92.0},"Non-Live Parallel AST":{"raw":"89.50%","value":89.5},"Non-Live Parallel Multiple AST":{"raw":"86.50%","value":86.5},"Non-Live Simple AST":{"raw":"74.08%","value":74.08},"Relevance Detection":{"raw":"93.75%","value":93.75},"Web Search Acc":{"raw":"58.00%","value":58.0},"Web Search Base":{"raw":"64.00%","value":64.0},"Web Search No Snippet":{"raw":"52.00%","value":52.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"10.0","Format Sensitivity Standard Deviation":"2.77","Irrelevance Detection":"67.00%","Live Acc":"76.02%","Live Multiple AST":"74.74%","Live Parallel AST":"87.50%","Live Parallel Multiple AST":"54.17%","Live Simple AST":"82.56%","Memory Acc":"44.09%","Memory KV":"46.45%","Memory Recursive Summarization":"39.35%","Memory Vector":"46.45%","Multi Turn Acc":"44.88%","Multi Turn Base":"55.00%","Multi Turn Long Context":"48.50%","Multi Turn Miss Func":"49.00%","Multi Turn Miss Param":"27.00%","Non-Live AST Acc":"85.52%","Non-Live Multiple AST":"92.00%","Non-Live Parallel AST":"89.50%","Non-Live Parallel Multiple AST":"86.50%","Non-Live Simple AST":"74.08%","Relevance Detection":"93.75%","Web Search Acc":"58.00%","Web Search Base":"64.00%","Web Search No Snippet":"52.00%"},"total_cost_raw":"57.75","total_cost_usd":57.75,"variant":"Thinking"},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":56.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-8e86739001363f13bd49","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-5586ae213879dae6f344b7b9","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":19,"rawValue":"54.12%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"671ad7005b9522d9d14de8d1613e3e515ee79bf0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":19,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=18;rank=19;model=DeepSeek-V3.2-Exp (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":19,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"5.83","latency_mean_s":5.83,"latency_p95_raw":"10.59","latency_p95_s":10.59,"latency_std_raw":"11.71","latency_std_s":11.71,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"MIT","model":"DeepSeek-V3.2-Exp","model_link":"https://api-docs.deepseek.com/news/news250528","organization":"DeepSeek","overall_accuracy":"54.12%","overall_accuracy_percent":54.12,"rank":19,"source_model":"DeepSeek-V3.2-Exp (FC)","source_row_index":18,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"93.18%","value":93.18},"Live Acc":{"raw":"53.66%","value":53.66},"Live Multiple AST":{"raw":"51.66%","value":51.66},"Live Parallel AST":{"raw":"25.00%","value":25.0},"Live Parallel Multiple AST":{"raw":"25.00%","value":25.0},"Live Simple AST":{"raw":"66.28%","value":66.28},"Memory Acc":{"raw":"54.19%","value":54.19},"Memory KV":{"raw":"41.94%","value":41.94},"Memory Recursive Summarization":{"raw":"59.35%","value":59.35},"Memory Vector":{"raw":"61.29%","value":61.29},"Multi Turn Acc":{"raw":"37.38%","value":37.38},"Multi Turn Base":{"raw":"41.50%","value":41.5},"Multi Turn Long Context":{"raw":"35.00%","value":35.0},"Multi Turn Miss Func":{"raw":"39.50%","value":39.5},"Multi Turn Miss Param":{"raw":"33.50%","value":33.5},"Non-Live AST Acc":{"raw":"34.85%","value":34.85},"Non-Live Multiple AST":{"raw":"74.00%","value":74.0},"Non-Live Parallel AST":{"raw":"15.00%","value":15.0},"Non-Live Parallel Multiple AST":{"raw":"12.50%","value":12.5},"Non-Live Simple AST":{"raw":"37.92%","value":37.92},"Relevance Detection":{"raw":"37.50%","value":37.5},"Web Search Acc":{"raw":"69.50%","value":69.5},"Web Search Base":{"raw":"80.00%","value":80.0},"Web Search No Snippet":{"raw":"59.00%","value":59.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"93.18%","Live Acc":"53.66%","Live Multiple AST":"51.66%","Live Parallel AST":"25.00%","Live Parallel Multiple AST":"25.00%","Live Simple AST":"66.28%","Memory Acc":"54.19%","Memory KV":"41.94%","Memory Recursive Summarization":"59.35%","Memory Vector":"61.29%","Multi Turn Acc":"37.38%","Multi Turn Base":"41.50%","Multi Turn Long Context":"35.00%","Multi Turn Miss Func":"39.50%","Multi Turn Miss Param":"33.50%","Non-Live AST Acc":"34.85%","Non-Live Multiple AST":"74.00%","Non-Live Parallel AST":"15.00%","Non-Live Parallel Multiple AST":"12.50%","Non-Live Simple AST":"37.92%","Relevance Detection":"37.50%","Web Search Acc":"69.50%","Web Search Base":"80.00%","Web Search No Snippet":"59.00%"},"total_cost_raw":"6.71","total_cost_usd":6.71,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":54.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-28b791bf737ccebcb27b","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-b1b7deaae0ec48a3f84286b4","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"BFCL exact dated model spelling.","registry":"model_aliases.json"},"mappingNote":"BFCL exact dated model spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"GPT-5 mini","modelRef":"GPT-5-mini-2025-08-07","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":17,"rawValue":"55.46%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c44a6f7210f4d8a44c5a098a6d998622f48ff03","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":17,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=16;rank=17;model=GPT-5-mini-2025-08-07 (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":17,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"8.32","latency_mean_s":8.32,"latency_p95_raw":"19.8","latency_p95_s":19.8,"latency_std_raw":"17.35","latency_std_s":17.35,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"GPT-5-mini-2025-08-07","model_link":"https://openai.com/index/introducing-gpt-5/","organization":"OpenAI","overall_accuracy":"55.46%","overall_accuracy_percent":55.46,"rank":17,"source_model":"GPT-5-mini-2025-08-07 (FC)","source_row_index":16,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"91.01%","value":91.01},"Live Acc":{"raw":"58.62%","value":58.62},"Live Multiple AST":{"raw":"58.02%","value":58.02},"Live Parallel AST":{"raw":"62.50%","value":62.5},"Live Parallel Multiple AST":{"raw":"45.83%","value":45.83},"Live Simple AST":{"raw":"62.02%","value":62.02},"Memory Acc":{"raw":"44.30%","value":44.3},"Memory KV":{"raw":"36.77%","value":36.77},"Memory Recursive Summarization":{"raw":"52.26%","value":52.26},"Memory Vector":{"raw":"43.87%","value":43.87},"Multi Turn Acc":{"raw":"27.50%","value":27.5},"Multi Turn Base":{"raw":"36.50%","value":36.5},"Multi Turn Long Context":{"raw":"33.00%","value":33.0},"Multi Turn Miss Func":{"raw":"17.00%","value":17.0},"Multi Turn Miss Param":{"raw":"23.50%","value":23.5},"Non-Live AST Acc":{"raw":"69.85%","value":69.85},"Non-Live Multiple AST":{"raw":"69.00%","value":69.0},"Non-Live Parallel AST":{"raw":"80.00%","value":80.0},"Non-Live Parallel Multiple AST":{"raw":"70.50%","value":70.5},"Non-Live Simple AST":{"raw":"59.92%","value":59.92},"Relevance Detection":{"raw":"62.50%","value":62.5},"Web Search Acc":{"raw":"82.00%","value":82.0},"Web Search Base":{"raw":"87.00%","value":87.0},"Web Search No Snippet":{"raw":"77.00%","value":77.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"91.01%","Live Acc":"58.62%","Live Multiple AST":"58.02%","Live Parallel AST":"62.50%","Live Parallel Multiple AST":"45.83%","Live Simple AST":"62.02%","Memory Acc":"44.30%","Memory KV":"36.77%","Memory Recursive Summarization":"52.26%","Memory Vector":"43.87%","Multi Turn Acc":"27.50%","Multi Turn Base":"36.50%","Multi Turn Long Context":"33.00%","Multi Turn Miss Func":"17.00%","Multi Turn Miss Param":"23.50%","Non-Live AST Acc":"69.85%","Non-Live Multiple AST":"69.00%","Non-Live Parallel AST":"80.00%","Non-Live Parallel Multiple AST":"70.50%","Non-Live Simple AST":"59.92%","Relevance Detection":"62.50%","Web Search Acc":"82.00%","Web Search Base":"87.00%","Web Search No Snippet":"77.00%"},"total_cost_raw":"22.18","total_cost_usd":22.18,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":55.46,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-91b21445218fa9ec47dd","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-ca94f63940fd97e4c9f4d82b","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"BFCL exact dated model spelling.","registry":"model_aliases.json"},"mappingNote":"BFCL exact dated model spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"GPT-5 mini","modelRef":"GPT-5-mini-2025-08-07","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":77,"rawValue":"27.83%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c44a6f7210f4d8a44c5a098a6d998622f48ff03","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":77,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=76;rank=77;model=GPT-5-mini-2025-08-07 (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":77,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"8.89","latency_mean_s":8.89,"latency_p95_raw":"19.72","latency_p95_s":19.72,"latency_std_raw":"11.08","latency_std_s":11.08,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"GPT-5-mini-2025-08-07","model_link":"https://openai.com/index/introducing-gpt-5/","organization":"OpenAI","overall_accuracy":"27.83%","overall_accuracy_percent":27.83,"rank":77,"source_model":"GPT-5-mini-2025-08-07 (Prompt)","source_row_index":76,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"16.0","value":16.0},"Format Sensitivity Standard Deviation":{"raw":"3.78","value":3.78},"Irrelevance Detection":{"raw":"55.71%","value":55.71},"Live Acc":{"raw":"62.55%","value":62.55},"Live Multiple AST":{"raw":"61.16%","value":61.16},"Live Parallel AST":{"raw":"75.00%","value":75.0},"Live Parallel Multiple AST":{"raw":"37.50%","value":37.5},"Live Simple AST":{"raw":"69.77%","value":69.77},"Memory Acc":{"raw":"29.25%","value":29.25},"Memory KV":{"raw":"19.35%","value":19.35},"Memory Recursive Summarization":{"raw":"38.71%","value":38.71},"Memory Vector":{"raw":"29.68%","value":29.68},"Multi Turn Acc":{"raw":"5.50%","value":5.5},"Multi Turn Base":{"raw":"5.50%","value":5.5},"Multi Turn Long Context":{"raw":"7.00%","value":7.0},"Multi Turn Miss Func":{"raw":"5.00%","value":5.0},"Multi Turn Miss Param":{"raw":"4.50%","value":4.5},"Non-Live AST Acc":{"raw":"68.04%","value":68.04},"Non-Live Multiple AST":{"raw":"72.50%","value":72.5},"Non-Live Parallel AST":{"raw":"71.50%","value":71.5},"Non-Live Parallel Multiple AST":{"raw":"69.00%","value":69.0},"Non-Live Simple AST":{"raw":"59.17%","value":59.17},"Relevance Detection":{"raw":"93.75%","value":93.75},"Web Search Acc":{"raw":"8.50%","value":8.5},"Web Search Base":{"raw":"11.00%","value":11.0},"Web Search No Snippet":{"raw":"6.00%","value":6.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"16.0","Format Sensitivity Standard Deviation":"3.78","Irrelevance Detection":"55.71%","Live Acc":"62.55%","Live Multiple AST":"61.16%","Live Parallel AST":"75.00%","Live Parallel Multiple AST":"37.50%","Live Simple AST":"69.77%","Memory Acc":"29.25%","Memory KV":"19.35%","Memory Recursive Summarization":"38.71%","Memory Vector":"29.68%","Multi Turn Acc":"5.50%","Multi Turn Base":"5.50%","Multi Turn Long Context":"7.00%","Multi Turn Miss Func":"5.00%","Multi Turn Miss Param":"4.50%","Non-Live AST Acc":"68.04%","Non-Live Multiple AST":"72.50%","Non-Live Parallel AST":"71.50%","Non-Live Parallel Multiple AST":"69.00%","Non-Live Simple AST":"59.17%","Relevance Detection":"93.75%","Web Search Acc":"8.50%","Web Search Base":"11.00%","Web Search No Snippet":"6.00%"},"total_cost_raw":"82.74","total_cost_usd":82.74,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":27.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-8a132f6bb460644d2948","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-15e5d8399c6f1b82d96bec8c","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"BFCL exact dated model spelling.","registry":"model_aliases.json"},"mappingNote":"BFCL exact dated model spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"GPT-5 nano","modelRef":"GPT-5-nano-2025-08-07","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":24,"rawValue":"51.45%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df6676bacd282bde0fae3be8d70a8644e83d2815","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":24,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=23;rank=24;model=GPT-5-nano-2025-08-07 (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":24,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"10.36","latency_mean_s":10.36,"latency_p95_raw":"23.56","latency_p95_s":23.56,"latency_std_raw":"10.37","latency_std_s":10.37,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"GPT-5-nano-2025-08-07","model_link":"https://openai.com/index/introducing-gpt-5/","organization":"OpenAI","overall_accuracy":"51.45%","overall_accuracy_percent":51.45,"rank":24,"source_model":"GPT-5-nano-2025-08-07 (FC)","source_row_index":23,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"89.10%","value":89.1},"Live Acc":{"raw":"59.44%","value":59.44},"Live Multiple AST":{"raw":"59.83%","value":59.83},"Live Parallel AST":{"raw":"50.00%","value":50.0},"Live Parallel Multiple AST":{"raw":"54.17%","value":54.17},"Live Simple AST":{"raw":"58.91%","value":58.91},"Memory Acc":{"raw":"24.73%","value":24.73},"Memory KV":{"raw":"18.06%","value":18.06},"Memory Recursive Summarization":{"raw":"29.03%","value":29.03},"Memory Vector":{"raw":"27.10%","value":27.1},"Multi Turn Acc":{"raw":"34.50%","value":34.5},"Multi Turn Base":{"raw":"44.00%","value":44.0},"Multi Turn Long Context":{"raw":"38.00%","value":38.0},"Multi Turn Miss Func":{"raw":"23.50%","value":23.5},"Multi Turn Miss Param":{"raw":"32.50%","value":32.5},"Non-Live AST Acc":{"raw":"68.00%","value":68.0},"Non-Live Multiple AST":{"raw":"64.50%","value":64.5},"Non-Live Parallel AST":{"raw":"79.00%","value":79.0},"Non-Live Parallel Multiple AST":{"raw":"71.50%","value":71.5},"Non-Live Simple AST":{"raw":"57.00%","value":57.0},"Relevance Detection":{"raw":"75.00%","value":75.0},"Web Search Acc":{"raw":"72.50%","value":72.5},"Web Search Base":{"raw":"74.00%","value":74.0},"Web Search No Snippet":{"raw":"71.00%","value":71.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"89.10%","Live Acc":"59.44%","Live Multiple AST":"59.83%","Live Parallel AST":"50.00%","Live Parallel Multiple AST":"54.17%","Live Simple AST":"58.91%","Memory Acc":"24.73%","Memory KV":"18.06%","Memory Recursive Summarization":"29.03%","Memory Vector":"27.10%","Multi Turn Acc":"34.50%","Multi Turn Base":"44.00%","Multi Turn Long Context":"38.00%","Multi Turn Miss Func":"23.50%","Multi Turn Miss Param":"32.50%","Non-Live AST Acc":"68.00%","Non-Live Multiple AST":"64.50%","Non-Live Parallel AST":"79.00%","Non-Live Parallel Multiple AST":"71.50%","Non-Live Simple AST":"57.00%","Relevance Detection":"75.00%","Web Search Acc":"72.50%","Web Search Base":"74.00%","Web Search No Snippet":"71.00%"},"total_cost_raw":"8.79","total_cost_usd":8.79,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":51.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-115956bcd886cdbc6354","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-66d7f2f32bf3bece91140771","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"BFCL exact dated model spelling.","registry":"model_aliases.json"},"mappingNote":"BFCL exact dated model spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"GPT-5 nano","modelRef":"GPT-5-nano-2025-08-07","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":79,"rawValue":"27.55%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df6676bacd282bde0fae3be8d70a8644e83d2815","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":79,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=78;rank=79;model=GPT-5-nano-2025-08-07 (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":79,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"10.67","latency_mean_s":10.67,"latency_p95_raw":"23.28","latency_p95_s":23.28,"latency_std_raw":"7.68","latency_std_s":7.68,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"GPT-5-nano-2025-08-07","model_link":"https://openai.com/index/introducing-gpt-5/","organization":"OpenAI","overall_accuracy":"27.55%","overall_accuracy_percent":27.55,"rank":79,"source_model":"GPT-5-nano-2025-08-07 (Prompt)","source_row_index":78,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"8.5","value":8.5},"Format Sensitivity Standard Deviation":{"raw":"2.57","value":2.57},"Irrelevance Detection":{"raw":"45.75%","value":45.75},"Live Acc":{"raw":"70.69%","value":70.69},"Live Multiple AST":{"raw":"69.71%","value":69.71},"Live Parallel AST":{"raw":"68.75%","value":68.75},"Live Parallel Multiple AST":{"raw":"54.17%","value":54.17},"Live Simple AST":{"raw":"76.36%","value":76.36},"Memory Acc":{"raw":"24.52%","value":24.52},"Memory KV":{"raw":"20.65%","value":20.65},"Memory Recursive Summarization":{"raw":"21.29%","value":21.29},"Memory Vector":{"raw":"31.61%","value":31.61},"Multi Turn Acc":{"raw":"0.75%","value":0.75},"Multi Turn Base":{"raw":"1.00%","value":1.0},"Multi Turn Long Context":{"raw":"1.00%","value":1.0},"Multi Turn Miss Func":{"raw":"1.00%","value":1.0},"Multi Turn Miss Param":{"raw":"0.00%","value":0.0},"Non-Live AST Acc":{"raw":"80.81%","value":80.81},"Non-Live Multiple AST":{"raw":"86.00%","value":86.0},"Non-Live Parallel AST":{"raw":"87.50%","value":87.5},"Non-Live Parallel Multiple AST":{"raw":"80.50%","value":80.5},"Non-Live Simple AST":{"raw":"69.25%","value":69.25},"Relevance Detection":{"raw":"93.75%","value":93.75},"Web Search Acc":{"raw":"13.50%","value":13.5},"Web Search Base":{"raw":"10.00%","value":10.0},"Web Search No Snippet":{"raw":"17.00%","value":17.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"8.5","Format Sensitivity Standard Deviation":"2.57","Irrelevance Detection":"45.75%","Live Acc":"70.69%","Live Multiple AST":"69.71%","Live Parallel AST":"68.75%","Live Parallel Multiple AST":"54.17%","Live Simple AST":"76.36%","Memory Acc":"24.52%","Memory KV":"20.65%","Memory Recursive Summarization":"21.29%","Memory Vector":"31.61%","Multi Turn Acc":"0.75%","Multi Turn Base":"1.00%","Multi Turn Long Context":"1.00%","Multi Turn Miss Func":"1.00%","Multi Turn Miss Param":"0.00%","Non-Live AST Acc":"80.81%","Non-Live Multiple AST":"86.00%","Non-Live Parallel AST":"87.50%","Non-Live Parallel Multiple AST":"80.50%","Non-Live Simple AST":"69.25%","Relevance Detection":"93.75%","Web Search Acc":"13.50%","Web Search Base":"10.00%","Web Search No Snippet":"17.00%"},"total_cost_raw":"21.47","total_cost_usd":21.47,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":27.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-c30919fbdb262a7857b8","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-bdea4dfd0cff4904f2ac7e5a","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"GPT-5.2","modelRef":"GPT-5.2-2025-12-11","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":16,"rawValue":"55.87%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7be40fd959ab047a1b92682ed8370abe6ce1c54","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":16,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=15;rank=16;model=GPT-5.2-2025-12-11 (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":16,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"2.23","latency_mean_s":2.23,"latency_p95_raw":"5.26","latency_p95_s":5.26,"latency_std_raw":"9.75","latency_std_s":9.75,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"GPT-5.2-2025-12-11","model_link":"https://openai.com/zh-Hans-CN/index/introducing-gpt-5-2/","organization":"OpenAI","overall_accuracy":"55.87%","overall_accuracy_percent":55.87,"rank":16,"source_model":"GPT-5.2-2025-12-11 (FC)","source_row_index":15,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"79.42%","value":79.42},"Live Acc":{"raw":"70.39%","value":70.39},"Live Multiple AST":{"raw":"70.37%","value":70.37},"Live Parallel AST":{"raw":"68.75%","value":68.75},"Live Parallel Multiple AST":{"raw":"58.33%","value":58.33},"Live Simple AST":{"raw":"71.71%","value":71.71},"Memory Acc":{"raw":"45.81%","value":45.81},"Memory KV":{"raw":"33.55%","value":33.55},"Memory Recursive Summarization":{"raw":"60.65%","value":60.65},"Memory Vector":{"raw":"43.23%","value":43.23},"Multi Turn Acc":{"raw":"28.12%","value":28.12},"Multi Turn Base":{"raw":"36.50%","value":36.5},"Multi Turn Long Context":{"raw":"30.50%","value":30.5},"Multi Turn Miss Func":{"raw":"18.00%","value":18.0},"Multi Turn Miss Param":{"raw":"27.50%","value":27.5},"Non-Live AST Acc":{"raw":"81.85%","value":81.85},"Non-Live Multiple AST":{"raw":"88.00%","value":88.0},"Non-Live Parallel AST":{"raw":"89.00%","value":89.0},"Non-Live Parallel Multiple AST":{"raw":"77.50%","value":77.5},"Non-Live Simple AST":{"raw":"72.92%","value":72.92},"Relevance Detection":{"raw":"75.00%","value":75.0},"Web Search Acc":{"raw":"75.50%","value":75.5},"Web Search Base":{"raw":"78.00%","value":78.0},"Web Search No Snippet":{"raw":"73.00%","value":73.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"79.42%","Live Acc":"70.39%","Live Multiple AST":"70.37%","Live Parallel AST":"68.75%","Live Parallel Multiple AST":"58.33%","Live Simple AST":"71.71%","Memory Acc":"45.81%","Memory KV":"33.55%","Memory Recursive Summarization":"60.65%","Memory Vector":"43.23%","Multi Turn Acc":"28.12%","Multi Turn Base":"36.50%","Multi Turn Long Context":"30.50%","Multi Turn Miss Func":"18.00%","Multi Turn Miss Param":"27.50%","Non-Live AST Acc":"81.85%","Non-Live Multiple AST":"88.00%","Non-Live Parallel AST":"89.00%","Non-Live Parallel Multiple AST":"77.50%","Non-Live Simple AST":"72.92%","Relevance Detection":"75.00%","Web Search Acc":"75.50%","Web Search Base":"78.00%","Web Search No Snippet":"73.00%"},"total_cost_raw":"85.65","total_cost_usd":85.65,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":55.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-355015f6267c5ac02d68","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-48710143ac56f4f2ae4452ee","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"GPT-5.2","modelRef":"GPT-5.2-2025-12-11","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","curated_model_alias","missing_source_date","mode_specific","official_csv"],"rank":38,"rawValue":"45.27%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7be40fd959ab047a1b92682ed8370abe6ce1c54","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":38,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=37;rank=38;model=GPT-5.2-2025-12-11 (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":38,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"4.21","latency_mean_s":4.21,"latency_p95_raw":"10.58","latency_p95_s":10.58,"latency_std_raw":"20.93","latency_std_s":20.93,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"GPT-5.2-2025-12-11","model_link":"https://openai.com/zh-Hans-CN/index/introducing-gpt-5-2/","organization":"OpenAI","overall_accuracy":"45.27%","overall_accuracy_percent":45.27,"rank":38,"source_model":"GPT-5.2-2025-12-11 (Prompt)","source_row_index":37,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"13.0","value":13.0},"Format Sensitivity Standard Deviation":{"raw":"3.25","value":3.25},"Irrelevance Detection":{"raw":"87.26%","value":87.26},"Live Acc":{"raw":"67.14%","value":67.14},"Live Multiple AST":{"raw":"64.58%","value":64.58},"Live Parallel AST":{"raw":"75.00%","value":75.0},"Live Parallel Multiple AST":{"raw":"58.33%","value":58.33},"Live Simple AST":{"raw":"77.91%","value":77.91},"Memory Acc":{"raw":"3.87%","value":3.87},"Memory KV":{"raw":"2.58%","value":2.58},"Memory Recursive Summarization":{"raw":"7.10%","value":7.1},"Memory Vector":{"raw":"1.94%","value":1.94},"Multi Turn Acc":{"raw":"43.75%","value":43.75},"Multi Turn Base":{"raw":"54.50%","value":54.5},"Multi Turn Long Context":{"raw":"46.50%","value":46.5},"Multi Turn Miss Func":{"raw":"40.50%","value":40.5},"Multi Turn Miss Param":{"raw":"33.50%","value":33.5},"Non-Live AST Acc":{"raw":"78.29%","value":78.29},"Non-Live Multiple AST":{"raw":"83.50%","value":83.5},"Non-Live Parallel AST":{"raw":"84.00%","value":84.0},"Non-Live Parallel Multiple AST":{"raw":"74.50%","value":74.5},"Non-Live Simple AST":{"raw":"71.17%","value":71.17},"Relevance Detection":{"raw":"75.00%","value":75.0},"Web Search Acc":{"raw":"40.50%","value":40.5},"Web Search Base":{"raw":"45.00%","value":45.0},"Web Search No Snippet":{"raw":"36.00%","value":36.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"13.0","Format Sensitivity Standard Deviation":"3.25","Irrelevance Detection":"87.26%","Live Acc":"67.14%","Live Multiple AST":"64.58%","Live Parallel AST":"75.00%","Live Parallel Multiple AST":"58.33%","Live Simple AST":"77.91%","Memory Acc":"3.87%","Memory KV":"2.58%","Memory Recursive Summarization":"7.10%","Memory Vector":"1.94%","Multi Turn Acc":"43.75%","Multi Turn Base":"54.50%","Multi Turn Long Context":"46.50%","Multi Turn Miss Func":"40.50%","Multi Turn Miss Param":"33.50%","Non-Live AST Acc":"78.29%","Non-Live Multiple AST":"83.50%","Non-Live Parallel AST":"84.00%","Non-Live Parallel Multiple AST":"74.50%","Non-Live Simple AST":"71.17%","Relevance Detection":"75.00%","Web Search Acc":"40.50%","Web Search Base":"45.00%","Web Search No Snippet":"36.00%"},"total_cost_raw":"164.58","total_cost_usd":164.58,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":45.27,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-6670f9c672b6e345631e","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-4ca0cc84048f1d321a8124b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Gemini 2.5 Flash","modelRef":"Gemini-2.5-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","missing_source_date","missing_submetrics","mode_specific","official_csv"],"rank":15,"rawValue":"56.24%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f4fa72de680a2663bed00cfc644a343a9fef54d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":15,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=14;rank=15;model=Gemini-2.5-Flash (FC);mode=FC","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":15,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"native_fc","calling_mode_label":"FC","evaluator_commit":"f7cf735","latency_mean_raw":"2.99","latency_mean_s":2.99,"latency_p95_raw":"5.62","latency_p95_s":5.62,"latency_std_raw":"9.22","latency_std_s":9.22,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Gemini-2.5-Flash","model_link":"https://deepmind.google/technologies/gemini/flash/","organization":"Google","overall_accuracy":"56.24%","overall_accuracy_percent":56.24,"rank":15,"source_model":"Gemini-2.5-Flash (FC)","source_row_index":14,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"N/A","value":null},"Format Sensitivity Standard Deviation":{"raw":"N/A","value":null},"Irrelevance Detection":{"raw":"93.67%","value":93.67},"Live Acc":{"raw":"74.39%","value":74.39},"Live Multiple AST":{"raw":"71.70%","value":71.7},"Live Parallel AST":{"raw":"81.25%","value":81.25},"Live Parallel Multiple AST":{"raw":"70.83%","value":70.83},"Live Simple AST":{"raw":"85.27%","value":85.27},"Memory Acc":{"raw":"41.29%","value":41.29},"Memory KV":{"raw":"19.35%","value":19.35},"Memory Recursive Summarization":{"raw":"54.19%","value":54.19},"Memory Vector":{"raw":"50.32%","value":50.32},"Multi Turn Acc":{"raw":"36.25%","value":36.25},"Multi Turn Base":{"raw":"41.50%","value":41.5},"Multi Turn Long Context":{"raw":"35.50%","value":35.5},"Multi Turn Miss Func":{"raw":"36.00%","value":36.0},"Multi Turn Miss Param":{"raw":"32.00%","value":32.0},"Non-Live AST Acc":{"raw":"84.96%","value":84.96},"Non-Live Multiple AST":{"raw":"92.00%","value":92.0},"Non-Live Parallel AST":{"raw":"94.00%","value":94.0},"Non-Live Parallel Multiple AST":{"raw":"79.50%","value":79.5},"Non-Live Simple AST":{"raw":"74.33%","value":74.33},"Relevance Detection":{"raw":"75.00%","value":75.0},"Web Search Acc":{"raw":"59.00%","value":59.0},"Web Search Base":{"raw":"59.00%","value":59.0},"Web Search No Snippet":{"raw":"59.00%","value":59.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"N/A","Format Sensitivity Standard Deviation":"N/A","Irrelevance Detection":"93.67%","Live Acc":"74.39%","Live Multiple AST":"71.70%","Live Parallel AST":"81.25%","Live Parallel Multiple AST":"70.83%","Live Simple AST":"85.27%","Memory Acc":"41.29%","Memory KV":"19.35%","Memory Recursive Summarization":"54.19%","Memory Vector":"50.32%","Multi Turn Acc":"36.25%","Multi Turn Base":"41.50%","Multi Turn Long Context":"35.50%","Multi Turn Miss Func":"36.00%","Multi Turn Miss Param":"32.00%","Non-Live AST Acc":"84.96%","Non-Live Multiple AST":"92.00%","Non-Live Parallel AST":"94.00%","Non-Live Parallel Multiple AST":"79.50%","Non-Live Simple AST":"74.33%","Relevance Detection":"75.00%","Web Search Acc":"59.00%","Web Search Base":"59.00%","Web Search No Snippet":"59.00%"},"total_cost_raw":"26.36","total_cost_usd":26.36,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","native_fc"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":56.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"bfcl","benchmarkName":"Berkeley Function Calling Leaderboard","benchmarkVersion":"BFCL-V4","benchmarkVersionHint":null,"benchmarkVersionId":"bfcl@v4","benchmarkVersionStatus":"source_reported","candidateId":"cand-src-bfcl-a999ba3ebc2e50bb4715","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gorilla.cs.berkeley.edu/leaderboard","harness":"bfcl-eval","harnessId":"bfcl-eval","httpStatus":200,"id":"pub-8c36bfe4115a81beef1af1b0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"accuracy","modelName":"Gemini 2.5 Flash","modelRef":"Gemini-2.5-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","protocol":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","commit":"f7cf735","commit_url":"https://github.com/ShishirPatil/gorilla/commit/f7cf7359b7ac615a0b294831c5ba2bc95ee4a000","evaluator":"bfcl","evaluator_commit":"f7cf735","harness":"bfcl-eval","leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","subject_type":"system","variant":null},"publishedAt":null,"qualityFlags":["benchmark_v4","missing_source_date","mode_specific","official_csv"],"rank":26,"rawValue":"50.90%","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f4fa72de680a2663bed00cfc644a343a9fef54d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-bfcl/candidates.jsonl","line":26,"payloadSha256":"ede5081a24bc5ac7f9b3b09a52af6bcfcd35aa0f1709239fed979db77a8159a6","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","sourceId":"src-bfcl","sourceLabel":"Berkeley Function Calling Leaderboard","sourceLocator":"csv-row=25;rank=26;model=Gemini-2.5-Flash (Prompt);mode=Prompt","sourceMetricId":"accuracy","sourcePageUrl":"https://gorilla.cs.berkeley.edu/leaderboard","sourceRow":{"artifact":"src-bfcl/candidates.jsonl","line":26,"metadata":{"benchmark_version":"BFCL-V4","benchmark_version_id":"bfcl@v4","calling_mode":"prompt","calling_mode_label":"Prompt","evaluator_commit":"f7cf735","latency_mean_raw":"3.18","latency_mean_s":3.18,"latency_p95_raw":"6.09","latency_p95_s":6.09,"latency_std_raw":"4.44","latency_std_s":4.44,"leaderboard_url":"https://gorilla.cs.berkeley.edu/leaderboard","license":"Proprietary","model":"Gemini-2.5-Flash","model_link":"https://deepmind.google/technologies/gemini/flash/","organization":"Google","overall_accuracy":"50.90%","overall_accuracy_percent":50.9,"rank":26,"source_model":"Gemini-2.5-Flash (Prompt)","source_row_index":25,"source_status":"official_published_csv","submetrics":{"Format Sensitivity Max Delta":{"raw":"9.0","value":9.0},"Format Sensitivity Standard Deviation":{"raw":"2.45","value":2.45},"Irrelevance Detection":{"raw":"91.09%","value":91.09},"Live Acc":{"raw":"78.16%","value":78.16},"Live Multiple AST":{"raw":"75.97%","value":75.97},"Live Parallel AST":{"raw":"81.25%","value":81.25},"Live Parallel Multiple AST":{"raw":"75.00%","value":75.0},"Live Simple AST":{"raw":"87.21%","value":87.21},"Memory Acc":{"raw":"38.71%","value":38.71},"Memory KV":{"raw":"13.55%","value":13.55},"Memory Recursive Summarization":{"raw":"55.48%","value":55.48},"Memory Vector":{"raw":"47.10%","value":47.1},"Multi Turn Acc":{"raw":"16.75%","value":16.75},"Multi Turn Base":{"raw":"14.50%","value":14.5},"Multi Turn Long Context":{"raw":"18.50%","value":18.5},"Multi Turn Miss Func":{"raw":"16.50%","value":16.5},"Multi Turn Miss Param":{"raw":"17.50%","value":17.5},"Non-Live AST Acc":{"raw":"88.08%","value":88.08},"Non-Live Multiple AST":{"raw":"91.50%","value":91.5},"Non-Live Parallel AST":{"raw":"96.00%","value":96.0},"Non-Live Parallel Multiple AST":{"raw":"87.50%","value":87.5},"Non-Live Simple AST":{"raw":"77.33%","value":77.33},"Relevance Detection":{"raw":"62.50%","value":62.5},"Web Search Acc":{"raw":"62.00%","value":62.0},"Web Search Base":{"raw":"60.00%","value":60.0},"Web Search No Snippet":{"raw":"64.00%","value":64.0}},"submetrics_raw":{"Format Sensitivity Max Delta":"9.0","Format Sensitivity Standard Deviation":"2.45","Irrelevance Detection":"91.09%","Live Acc":"78.16%","Live Multiple AST":"75.97%","Live Parallel AST":"81.25%","Live Parallel Multiple AST":"75.00%","Live Simple AST":"87.21%","Memory Acc":"38.71%","Memory KV":"13.55%","Memory Recursive Summarization":"55.48%","Memory Vector":"47.10%","Multi Turn Acc":"16.75%","Multi Turn Base":"14.50%","Multi Turn Long Context":"18.50%","Multi Turn Miss Func":"16.50%","Multi Turn Miss Param":"17.50%","Non-Live AST Acc":"88.08%","Non-Live Multiple AST":"91.50%","Non-Live Parallel AST":"96.00%","Non-Live Parallel Multiple AST":"87.50%","Non-Live Simple AST":"77.33%","Relevance Detection":"62.50%","Web Search Acc":"62.00%","Web Search Base":"60.00%","Web Search No Snippet":"64.00%"},"total_cost_raw":"33.45","total_cost_usd":33.45,"variant":null},"sourceStatus":"official_published_csv","source_flags":["official_csv","prompt"]},"sourceSubjectType":"system","sourceUrl":"https://gorilla.cs.berkeley.edu/data_overall.csv","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=tool-use-eval"],"subjectType":"system","unit":"percent","value":50.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3ca06ce3373a9799ecf5","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08f95142e630a2ecaab729f3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"804.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8978999224a3e80e1a70e0c20325dbe85a66444","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":144,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=58;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":144,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":804.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-09-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-82ea1beaa126cd70ed6c","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b596f19ea5bc0fe806bfe64","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V3.1 Terminus","modelRef":"DeepSeek-V3.1-Terminus","observedAt":"2025-09-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"745.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da7c03da581a40b27c348f2ed062f0d42142915f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":154,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=68;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":154,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":745.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-37a5f2102f0872e5eb1e","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e679a0da6147db5d7e002ab1","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiniMax M2.1","modelRef":"MiniMax-M2.1","observedAt":"2025-12-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"623.83","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c989abc666312b606d15a7d859a8960c864a057b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":166,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=80;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":166,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"MiniMax","release_date":"2025-12-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":623.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b36fe6e3da737d3c7cf","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f5cd3d2031f255fbcaf84bf0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"618.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6623ab4b273bb96597f34aa84d3da22a87f07849","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":168,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=82;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":168,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":618.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0e86895efc4946e871f9","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f930522a1a8c26f73ae5f5a0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"599.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fa636085698d3e98077fbd447ac66b600f7a137","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":169,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=83;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":169,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":599.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-049372e1e443e410b92a","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa44259f1d5c953f1a607afa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"640.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05f900d03f1237631f2b06f8a772fe00c4a9cf91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":165,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=79;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":165,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":640.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-99fad153570f18cd4d2e","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-620bebf5b34932a105dbb5f9","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"461.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32eeebd39ce8f598792c788d4c568b59d6b9f1d0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":178,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=92;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":178,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":461.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-225a95fa7df6b387d5c1","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fb04f7a6d10a12372ef25b0a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2041.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0600e2f08b3cda9a9aeb5cd69e8a65998ebef366","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":90,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=4;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":90,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2041.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a906e4b192d7431fe6b1","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f2cd9e6ed2607e6c94c7a5d4","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"653.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"745bfb6d06238ab510fe7fa1dc30bfc325d9c52c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":164,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=78;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":164,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":653.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6dd4e4f64b2972ca2088","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8331184009343cb360f560d6","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1025.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26b3bd58ebfb3f96a09290092e895d903f5761fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":126,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=40;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":126,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1025.38,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7ba0ffd61e8c0397b418","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2ae1dd5a7b0057c189bde00","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"996.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"317e456bf5d84b87316a81f55698bbc0a69e5c14","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":130,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=44;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":130,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":996.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d372eb6f5463eaa69dfd","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-183796f64e699f8a1babe783","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1323.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77a2a08f30081d8ffd0ebc8790cf2dda310363f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":106,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=20;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":106,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1323.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-80b185b6b6d741bd7150","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd23e15b654e2c6d61a6037f","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1563.83","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8e9ce6566a78452730f139bd25ec11bcdacbf91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":97,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=11;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":97,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1563.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-989b2e6b08ff531051aa","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7a60589d5bc488ad6344e1b5","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1411.83","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8e9ce6566a78452730f139bd25ec11bcdacbf91","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":102,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=16;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":102,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1411.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fa379be2549b43ea243e","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-628f5795d91a8d95ebc22b5d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2164.57","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0eda3b8e42751cbc6368d18148afd7da28ee1ba9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":89,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=3;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":89,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2164.57,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ea0c4d7b6a5b518064b","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b389b3856bd32af288bfbdfb","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"796.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c9832b82af0795845d4e767ddb5f2de5070aae6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":147,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=61;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":147,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":796.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec56200017579f5b68f4","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-28801e1a76bc7b0eceea7b56","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1327.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f89f20b2a336993af7004f3e04beecc90ee4614","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":105,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=19;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":105,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1327.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1638c1ae265258ef3142","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d63d09f0680cff8abf3aa28e","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1463.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a1eb8ffc4083c9c754557ffbcc18f27ceaf75b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":101,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=15;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":101,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1463.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-40bbd15244cb08457b74","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cdfd4be1a5220fa234805f5b","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1306.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73977078a33a64640b86e995d096db3f84ea430d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":109,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=23;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":109,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1306.08,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ee190fe348e98ba7513","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-296460808526f01e333b8304","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"678.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73977078a33a64640b86e995d096db3f84ea430d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":159,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=73;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":159,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":678.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bb8e7d07a4e7df0401a2","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1875f25bc38f30549bb69e49","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"324.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73977078a33a64640b86e995d096db3f84ea430d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":189,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=103;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":189,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":324.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cd67941d80ad59bd5519","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3bf4cef52b306abe0e0b18fb","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1403.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"639c4ffe66b53ae14dd9ff8cd1877a3e9a20698a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":103,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=17;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":103,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1403.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4e20679bc9f2d288d695","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-14f794936dc5fb2b603d23dc","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1006.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a23e0351fa740da204b2c626dcb930679f33a966","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":128,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=42;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":128,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1006.08,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9aa4967366e43f7f89e6","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-958f26b4a946d4720ea1be11","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"521.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a23e0351fa740da204b2c626dcb930679f33a966","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":175,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=89;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":175,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":521.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3e45f92d7af3d881052e","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba86b78a196c859cd8066e85","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"661.88","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f19b3bb484d15c841568f360f2a0ef6a3542ae49","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":162,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=76;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":162,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":661.88,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8856bae11bc9699e0800","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f1cd5de990ee912a25ba2a09","mappingCandidates":["google/gemini-2.5-pro@2025-06-05"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro_32K","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"785.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d455addf7f8903e0947957c3276c7b35a1f48095","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":149,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=63;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":149,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":785.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6682be0ce8e8a1388d12","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c9543e471986aa1dad45248","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1160.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"75de90b16581b4802da03d089b56915d7c8b52b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":120,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=34;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":120,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1160.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-408436c284a93c7eb8ff","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2868c2583ec1b30bc82a12c2","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"765.27","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dbdb4d370a57a0ca31a61c1a9b334f3e03095207","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":152,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=66;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":152,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":765.27,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-be04d6526cf6a3c9f112","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-66e06f179645e8ef8b0cb754","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"911.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"064737fe79416ea35f6bc249bf034aca88e25515","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":136,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=50;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":136,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":911.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2549487f8aeaa6a29a08","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d13525bb65774c0d038efbb5","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"715.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2de60aef19a8227bf1248d26707a0d79a5548bb7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":157,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=71;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":157,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":715.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-520b9ede4374a3e441c5","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cb580ff1412486cc5c78ff78","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"904.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2510a6e74c588e06fc4ec17e1fd7053fba15a80a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":137,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=51;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":137,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":904.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-56f32c272a7d8e810f18","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-43a7c18e6fcbd98305050800","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"399.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5ae14a5562cb21889b066aa5cc06334bb5cf370","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":181,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=95;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":181,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":399.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-595fe3394741342ec9c5","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b8225008c03fd9c5ea71526a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"765.62","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05d8d1c68600a8988f7192612cfa69c0571df709","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":151,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=65;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":151,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":765.62,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8547c274408a280dd2d3","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0801b90c67e84be86a1fd0af","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"887.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90971616902afcfe7bf43f2501a41a001831943f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":139,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=53;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":139,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":887.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cb6aea0a9ad541e38e41","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-59fea734a1f5ba152458c2c4","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GLM-5.2","modelRef":"glm-5.2_high","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1046.95","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4980bd2e9b6945f622bc0176ff035d793d00f2f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":125,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=39;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":125,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1046.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a1f8161ad5d05b71004","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cbfab6bbff361085579f89c9","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1010.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4980bd2e9b6945f622bc0176ff035d793d00f2f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":127,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=41;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":127,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1010.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ed2257e8f9511a3adca","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd59b8065013c0cc8f017428","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GLM-5.3","modelRef":"glm-5.3_high","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1317.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d222cf8a6e5922988ef39d28e0c871e31f447a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":107,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=21;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":107,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1317.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4293dd8ae40672f85cc7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-170514541587b3da76b11d49","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1162.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5da557196bf86d0786374b38d46a7094f42f4d6c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":119,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=33;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":119,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1162.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ae1c868ce025e159c83a","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b5a9a27b44ef7d18f95de82c","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"807.65","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5da557196bf86d0786374b38d46a7094f42f4d6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":143,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=57;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":143,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":807.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-204135d6a004e9d8ea46","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-119bc82c8c548ee4db7e4978","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"799.77","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f78de423c51ff1ea7b1c301214bb700f4895b342","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":145,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=59;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":145,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":799.77,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5248fb3fdec2e0333655","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-420059a80d99bf05b82c5304","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"718.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6490743ecdb2863fa1e3ad7847a94fb64d045ac0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":156,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=70;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":156,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":718.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-770a3e42f6ed2c650abc","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6c7e726f95f1c7ccb08d0c7f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1192.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"540c867495c5e9abfa6517d6dc62ba8cc6e19cf1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":115,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=29;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":115,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1192.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc5157c97fa114e35df1","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0d5f625fa7934080d9370263","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-11-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1244.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f1dfeccfa193c1141550f7d792b7e1c011e84b53","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":113,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=27;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":113,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1244.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ff44a7b52fd5972f37d","canonicalModelId":"openai/gpt-5.1-codex-max@2025-11-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-abc0d98495abc8bb7e4ffeee","mappingCandidates":["openai/gpt-5.1-codex-max@2025-11-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Max release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Max release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.1 Codex Max","modelRef":"gpt-5.1-codex-max","observedAt":"2025-11-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1208.83","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"444ecf0587dddc20fdac7dd781e3eb496eedf2e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":114,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=28;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":114,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1208.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0d9009bcd9b263b407b0","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-861986abe64e9dd746408fca","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1293.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90ef8331b294b465f305bbab6a00d769664b78e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":111,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=25;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":111,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1293.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aaef89880730d4caa116","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-765a0b6cbff72211b05f0908","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1249.83","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90ef8331b294b465f305bbab6a00d769664b78e1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":112,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=26;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":112,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1249.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6b8f8c2f8f4e7be360ba","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ee051ca13e374c249f70990","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2025-12-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1299.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"242c07caba4018b4f6f5fdb0462c77cd88aa40a6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":110,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=24;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":110,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1299.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-58df32cb0278e86c6806","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1d1e8473403607f4293b4cc9","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex_xhigh","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1655.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0561910b4778528c187e0528002a8fb988318e66","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":94,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=8;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":94,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1655.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-44a43308373011e91fe9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-87b23ab32fa9ff65fa3be109","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1607.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68d8937ae90ada5da0d1e2a6ec2b6d72490afe87","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":95,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=9;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":95,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1607.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7271666c4a21d145b503","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eafed19fdfecca2737dd73d1","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1520.72","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68d8937ae90ada5da0d1e2a6ec2b6d72490afe87","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":99,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=13;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":99,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1520.72,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-57c6c87317bb4876be20","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ce9d043d678154e297d2d77","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1086.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68d8937ae90ada5da0d1e2a6ec2b6d72490afe87","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":124,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=38;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":124,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1086.03,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5692f57f30c7f1e582ec","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2d95193fcfa912ab56c73fd","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1188.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"79574ef52c8b5a30542b97aad7f22968b5bbcb92","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":117,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=31;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":117,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1188.58,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca184db72f724908c8ba","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1ec17e7cfdd87e56b55bfd58","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1004.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1565748c1b12d86b65684ae8385489d5c189e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":129,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=43;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":129,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1004.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4603f06cc7db9197d065","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e388e9153aebf944e13e95df","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1589.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"646eb391c5eeea1eecee01f5594fc981f6f2852c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":96,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=10;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":96,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1589.38,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4a6c7be2a9df03dc2798","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-27487a09e289be8dc4b40a56","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1127.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"646eb391c5eeea1eecee01f5594fc981f6f2852c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":122,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=36;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":122,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1127.58,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7fd5e201c3cb4b7a0325","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bae240ff8814622c718f7c79","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1942.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"646eb391c5eeea1eecee01f5594fc981f6f2852c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":92,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=6;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":92,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1942.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c006c1f4d7cda73987a8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-081230318beba6136e374c92","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1667.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b5a5d192ff22f9bfe1db8641d8257f2ef8b825e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":93,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=7;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":93,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1667.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75f5bcb60b363dde13ad","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c9e70fb91a898aa6102d52a","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2176.88","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"040796a48d1cef789ed81eb640f33a5f7f1e7ef5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":88,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=2;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":88,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2176.88,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d8ffc7a55cf171f6ce70","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e9418e1af546fa1a58b2d564","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1951.42","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3565b8b14cba5b5a2d257e87f35e1830593e8e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":91,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=5;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":91,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1951.42,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1f7b7c9e125eb25dd4b2","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c35b155c6761f7cceb4ab87","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1309.28","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e3c8f79de6a4fc5c6ea10765482feb943a00ac6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":108,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=22;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":108,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1309.28,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-756a209ed13168ecb2a7","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-58c4a65def19878e4e832b4b","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1508.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"56ae6f9cc08b728db714efa8a4628d4abdcaac09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":100,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=14;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":100,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1508.08,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8551c090962bf231f07b","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5f4d900ae4977e97e1068000","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"821.65","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c645a108add40fefdde1ff8efdc6a0147af413e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":142,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=56;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":142,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":821.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8f3d5af30a86d0566cb7","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c8296981921bc361026c3eb3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1092.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2452017e9bceb7c1b661636a5b5a05126363f06d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":123,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=37;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":123,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1092.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e4340fb7c079953005b4","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fa861330e75c2fa6bcc14c07","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"886.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"726628c970e7ce6bc40804dd308a64d2fd1bdd2c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":140,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=54;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":140,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":886.23,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3b67571d49eee589915f","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8decfe7c72269457ea7bc129","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1524.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ebc5f3f5d183dc72e85c32f98b1b59c639fbab34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":98,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=12;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":98,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1524.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-ec7808decb0ac2e0c95f","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1eb273294e8cd0b336a4f9c4","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_unverified"],"rank":null,"rawValue":"737.95","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"931a1b0cc0fcae96c26007a0fd4fbab13c6fab93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":155,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=69;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":155,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":737.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-08dff227b53285f6cff5","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b5b8ae3a6689bb341ce1c44","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_unverified"],"rank":null,"rawValue":"785.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc8dfd57545c5038a4e7bf4b27a40307fa4b71cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":150,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=64;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":150,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":785.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-485f5abf949bfee1bdef","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a5431afe458a6a55601760c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["unit_unverified"],"rank":null,"rawValue":"513.95","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8decc1650d466bcd2f9af8b960ad923e945fdc8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":176,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=90;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":176,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":513.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-03007c8f9da624b47c5f","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-487e23487eb0211e4c172a58","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["unit_unverified"],"rank":null,"rawValue":"899.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cebeb5d902932bc7b8d9944f5c65a8299789e12","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":138,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=52;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":138,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":899.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d5ebd2e66866a0618d5","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4bf6951622c9f7c88ea95cd1","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27B","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"349.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"016cf0957243ed111f9a122cb30a1b8377ca5083","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":184,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=98;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":184,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":349.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-581df3f3abd543250b63","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c184cb5c6866aeeef28c9685","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"221.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d67d567d395ab116437d68eb4d93ec443572c8d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":193,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=107;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":193,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":221.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7482cc001d2c2473c870","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eda775488ae4ac62e384d36a","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"621.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e40a334c775e981f519c37225a1093f0a9d6010f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":167,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=81;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":167,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":621.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ece220d6da9aeb061d5b","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6b73981ff3aa527b3270bb59","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"326.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26d7f168688bc2a0b41c7ff9b27542f56a4134dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":187,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=101;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":187,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":326.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fa8b8032e00ee44888b9","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e22b349a91a118c65ca4ecc9","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"670.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"620278f259e710eeba2e01b1afae1e0f43639970","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":161,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=75;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":161,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":670.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-ale_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d260b1739da6ac82f72d","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e59c2fe2e06a031eb3ddecc","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Performance","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1189.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fdd96da92931e06a0ff80f080f717e605c2eac38","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":116,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"ale_bench_external.csv:row=30;column=Performance","sourceMetricId":"Performance","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":116,"metadata":{"epoch_file":"ale_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1189.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-091aa02a805c37d0b314","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-34d6ad7daeb2cda6eba2f07a","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.44","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3d8d26ab09f5f6c62834885305fc97c9c531548","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":210,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=14;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":210,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8f2036ffbce2f7da39a0","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53b8e0181f0f4ebe8a15654e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1247ddb14d11d32e15477a170aa2f7fdb7542d52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":214,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=18;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":214,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2a1ed6841579ab271fcd","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2b0532399b00694580a43732","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.77","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e996bf9777b69a0487564632e6dceabca500146b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":202,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":202,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.77,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b316ba6e9334b94fbb3c","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c0318621c099e83bc42fe7b6","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22e3f648161c130ce1b1437b687801f6b3fe12ef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":209,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=13;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":209,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7cfdbefd6b66e2277adc","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-542ccd20a1f93474d6cc8d15","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"18858edb43d95836e6ed836061e331c775d699f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":207,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":207,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-614dac8b36e135f49043","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-621d5ef0499060ba2ecd0668","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.51","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60f585fb5403c3bed1517cc490766313eb40d691","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":208,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":208,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.51,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6a3977017483ddbb4a53","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-16e40cfb28640bfa1e50ad29","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a45bbdbb94bac7cd9d91d32f832f8cf867ae371f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":199,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":199,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4dd66eed615e1def035c","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-47b5ab1ff7400f6c8ac2ae3d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6965d3575ea567d923b644a05643812cecdb11d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":205,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":205,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e0d91d982d71b3498312","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-70f182c1d468d887694f64b3","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffebc1d26086995e9c682af65f964a623303053b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":212,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=16;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":212,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.38,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a9ab126580c305d94209","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-699e423858371a1a01f008b9","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a9ada4ec58d4df3c46650e53c0d3d9e139bfb81","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":198,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":198,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-algotune_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e192fc901c4a1dcef203","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-72c68628ba670407c9e94ce2","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1.85","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"17d42db2089cf7db341f7565b92fb0d1dfeecaa7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":200,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"algotune_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":200,"metadata":{"epoch_file":"algotune_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a9e3e31d405724567d6a","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b47d1b696cda8796a1f80b01","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.062000000000000006","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e81841fba7c6abb7ae571f9a550756135f7356b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":270,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=56;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":270,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.200000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d8d37ec85b0477876dfb","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be6749e8bf13b8e9b97142ae","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30de684a4795ab7b865672a0af8ed78644a63027","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":216,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=2;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":216,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ec8ac7e643dcd7a71ef","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-76c5f3957158ceda041f0e7e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.08900000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d4224fc8216606383b923ef631363ea24ceeb96","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":265,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=51;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":265,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7a6333de6e4871f17e5c","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-25aa03c73e74364f3458c7aa","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.207","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be73304c616624a3fd357648d0ca1f08b131c762","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":248,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=34;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":248,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f39479ba257b0f47700","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bd317b900bdc406de9d23557","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.321","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fb1a0bcdabe0b71e81c87176a2a4870cbf8c24e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":237,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=23;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":237,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4871a7c314cda0a59a34","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f8eb966756fb3bf96f86501","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.324","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fb1a0bcdabe0b71e81c87176a2a4870cbf8c24e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":236,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=22;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":236,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-85d613bbe6ea96c3a5d6","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ee9e8ee5dda8c499ac63c98","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.339","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ebd36571bb300f97f70b95e467094035a7767e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":233,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=19;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":233,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.900000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0ddfb60e88fc50876a54","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e8ccf2fff85b759c940d33e3","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.425","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"791b5aa51bda2b97debe2c18a1f8408b88d03977","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":218,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=4;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":218,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ec97a2fc50eb0de0b70","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f88d8806baf4ac06df5f44f","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.435","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6c3ec97421745d7bb8a4f93b8c55da18ed92530b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":217,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=3;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":217,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b903121d48f45f678e4e","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a8111cb9264d11d9d5bf8f3","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.237","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"426d8ce6a8ae3b50adc584caf200066b9e2cfe14","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":244,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=30;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":244,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b7ccfd436c9e6c0040f5","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-90e256cc9318eaaf06bf470c","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.325","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7df66796fe40dc076e37ab17f620df3b7e357601","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":235,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=21;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":235,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e7a56626e74503c73141","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-46deb19c23cafd2a4124b65b","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6e02f4d906d575868f7399a856cce73dcd1aaa29","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":268,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=54;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":268,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-128e33d6463382815f22","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-219dceba4276a18945482690","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.018000000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6d8ae4b79079dec01c046f4b18be331dab62d31","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":275,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=61;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":275,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.8000000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-696b0197877d67279b13","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2fd7affb7d4fe8822fa5214","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.066","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4eae9339afe503aeeb1ab105dea6aa7c8761255","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":269,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=55;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":269,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.6000000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-942b4b4fdaecde84014d","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a1d2b7557e476c4b734d5cbf","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.335","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8b2edcbac35a01bd4063806d6031034b2040113","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":234,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=20;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":234,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-be1b4b793e6d524668a6","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-51e1e65f3dc74278a2bb550a","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.087","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02f197b53b5364bc7cad9bbdcbcfb108904d9931","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":267,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=53;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":267,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d956af9dc3f139f9e2f2","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-89fa49a1b66404c7d6dea15d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.172","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"544922ac51bfa3a72bd2c49578dc8c323de4148e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":256,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=42;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":256,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-db89064741b2df8aa02b","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-92b76508abeb5b0774bb57c8","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35600000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f52b0c56d39fb8ad0f4092a5e46c071ac3b9d6f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":229,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=15;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":229,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d6e17de5b5c5b7b5224f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-591469dcb64a4ef817a738cf","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.18300000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5253ecdc437f5f6b19992cd5e420309d6b1f88e2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":251,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=37;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":251,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e94661c15f125e33654","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8b880bba849df0b39e58cdea","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.18300000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5253ecdc437f5f6b19992cd5e420309d6b1f88e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":252,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=38;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":252,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cf779dc6cf50f541e3f7","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd59b5c08ddad55696875b0e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17500000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78852a483b1f2f50a133b1a98b4f04bec4a73ac4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":253,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=39;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":253,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bb761ac420214140c222","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba106e63b4a62fba4148fba2","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17500000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78852a483b1f2f50a133b1a98b4f04bec4a73ac4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":254,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=40;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":254,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c84455805b7fd049d12e","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0279c53be397a5efae1c95eb","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-11-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.207","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a783b85760742fe5e46c0a3419829b9df0b810c2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":247,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=33;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":247,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d884c2eb957324b48ef2","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2738290a213b5ab4afb0b4e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b03606dcf3c60f5e399a3835ce5cb114a7263224","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":245,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=31;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":245,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-15fc54745a21894fe5e1","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-72f838f3ed5af7a28fb6d6ff","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b03606dcf3c60f5e399a3835ce5cb114a7263224","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":246,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=32;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":246,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a9bbfe0abd59df67ae46","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-62a4e5b933f21e79e46af18c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.344","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b03606dcf3c60f5e399a3835ce5cb114a7263224","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":231,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=17;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":231,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3edc271e3cb06d73df62","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b4153f33032104cdf0ed37a","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2025-12-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.276","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6bfa5e1ab5186518877e3e4c3ecd005a03e1de5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":240,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=26;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":240,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04474dd9528f893c1309","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fc4b3f42892789c139ad94c0","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.318","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b438dc2cc9768245f2e46ee7ca4f5c6846061950","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":238,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=24;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":238,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cdea8bb7db64ec1f04c9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8cd169b0411110ac6ca1623b","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.349","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4f212bdbdaa8a293373df7c30080369321a4953","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":230,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=16;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":230,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4e345e5a33280bb9a18b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7eea6b9cf09843f9403c7a5a","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4f212bdbdaa8a293373df7c30080369321a4953","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":228,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=14;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":228,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c42ddd7b46f163735ea3","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7a54da3c445064fda56ae50","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.24600000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19be94408881c4bb683444951dec9da8e74cbaed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":242,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=28;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":242,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":24.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-202f59ddf2a00fe2fbc9","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa2e8459fe2c1bbca722faff","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16899999999999998","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f0a5ef1fd1321715f4cb3e16ce6427096809699","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":257,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=43;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":257,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8e4e2760ba389b9450a4","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-31166d2a718a0ce358877339","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.385","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"437fb7182a42cd68938a9748d4e8bef1a3169291","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":225,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=11;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":225,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f992e53207a9a61c9cf","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f4b3a113059f4a8bde205769","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.385","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"437fb7182a42cd68938a9748d4e8bef1a3169291","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":224,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=10;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":224,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f158e5df495ce9a386d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a8961600fe2d45c34365b7ad","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.399","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f61f268acbdef0b248720d416964e2fda313994","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":222,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=8;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":222,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.900000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d17408bd95bd5fc053b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a0662b31c62c3fee1a36980f","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_promax","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f61f268acbdef0b248720d416964e2fda313994","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":221,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=7;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":221,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-426b009f9a71ec2cd323","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db526d75da164588d62149fd","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.37700000000000006","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f61f268acbdef0b248720d416964e2fda313994","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":226,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=12;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":226,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bac5044475c3eadbc2e6","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f566e7de5ff2d452b5dce791","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.37700000000000006","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f61f268acbdef0b248720d416964e2fda313994","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":227,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=13;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":227,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4c6391427181091b1ed3","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-83aca449f74acdb850a4cec2","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.342","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"735855d723a15e7820cdd4e5dacc395d2cbf139b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":232,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=18;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":232,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f36c917de582c873d01","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f025d0651b750f7042a21075","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.41200000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bad19c615f33dc56f4e75ce19ba054171cb24f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":220,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=6;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":220,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5225b6cc0ebacb56b3a7","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-17a772396638a574f853c02a","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.14400000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"61b9b50ad1e0605106faf8acf230f2c8e38f6702","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":259,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=45;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":259,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.400000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c08814db4f252c4d7fee","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73e013504984321c2b94d76d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.189","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e780fb5bb3b8136ed5df9e6d8f3f5b33a8492672","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":250,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=36;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":250,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b35457a1e05c798c9bef","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-150873dde4b2fded68d184df","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.276","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d5dc3e662fd6dd074151837381a5700f84159e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":241,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=27;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":241,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f57c4a5caae6e1d80251","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e489572efc37fffeb3bb4dd","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.39299999999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13083577df59495d739627e589c51338e6fc899a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":223,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=9;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":223,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b8b0c3ba937e1eb9a75","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-31f05c3f601c000748fbe460","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.136","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"706a024d7dc3908ecf6ac4315cc48e3d6cefca19","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":260,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=46;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":260,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.600000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-apex_agents_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dbe7be23dcc33e518c76","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c63bf258832767ea3183a830","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1 score","modelName":"GLM-4.7","modelRef":"zai-org/glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.087","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"02f197b53b5364bc7cad9bbdcbcfb108904d9931","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":266,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"apex_agents_external.csv:row=52;column=Pass@1 score","sourceMetricId":"Pass@1 score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":266,"metadata":{"epoch_file":"apex_agents_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-05-28","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a854b3b81360a5c70417","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f16473033ab40169ff7469ef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0112","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de10320b84394833e9cf8fa50daf51bef06cde1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":459,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=183;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":459,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0112,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-12","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-904a5aa5d286845a76b8","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-07950142bbfd0e668bd6b767","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0486","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6cf75fa333fc91f0e78f3bede10bc1be3112b21","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":410,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=134;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":410,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0486,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-32a9cef148b8ee0b383d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30cccee9f26dd193839918b2","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef988f6fe74979757216e054fe449774d8d3c75d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":286,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":286,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-80834082d62a5e44978f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d3a4d07d2c7265d91f4f8743","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7680555555555555","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef988f6fe74979757216e054fe449774d8d3c75d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":302,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=26;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":302,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7680555555555555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ed9535f290f0ec6874c6","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd86b0c38c17b14273d6138c","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8916666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef988f6fe74979757216e054fe449774d8d3c75d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":282,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":282,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8916666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cab243eceb0900caa7b5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-945cab112d06431f79571f72","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_medium","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8250000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef988f6fe74979757216e054fe449774d8d3c75d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":299,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=23;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":299,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8250000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c210c4b4e159f3f296e3","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f0378b407623ccdf5868d0b9","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_xhigh","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8833333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef988f6fe74979757216e054fe449774d8d3c75d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":283,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":283,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8833333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ac7504d6cf9428343ed1","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7a0eca5252f5d51481006f19","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce7c13e685578c568ced4dcb672e58a2826aca70","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":456,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=180;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":456,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ae3237658e49bb53976b","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0c4ba39afe9a9368989bb1e0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_16K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0278","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce7c13e685578c568ced4dcb672e58a2826aca70","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":426,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=150;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":426,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0278,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c2ab179a5467321e000d","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-05ba3c2ff8efd1b562087d26","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_1K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce7c13e685578c568ced4dcb672e58a2826aca70","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":458,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=182;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":458,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0100df3a1f6ccad41a4b","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4b4b41a93933fdf6a8c5793b","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0403","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce7c13e685578c568ced4dcb672e58a2826aca70","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":417,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=141;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":417,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-94550a598972c7751aad","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d40233fbbe3383aa912ce96a","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_8K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0167","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce7c13e685578c568ced4dcb672e58a2826aca70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":447,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=171;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":447,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-83a6b1b02f2979395c4f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0fdc8ec58bfc123d8e361935","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.012700000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e52ef137eacee9984169b8d92c8c8658bfc4dd9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":454,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=178;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":454,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.012700000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f79b4a3e11ac2e5b969d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2f9ca887507818ee273f219e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0861","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e52ef137eacee9984169b8d92c8c8658bfc4dd9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":390,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=114;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":390,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0861,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bd02364aa7c2e391a89f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-610993504638601fe7c6a0aa","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_1K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e52ef137eacee9984169b8d92c8c8658bfc4dd9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":484,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=208;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":484,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ff4a811e9be0eafa236f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7d4debf90b688f510702ca44","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_8K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0452","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e52ef137eacee9984169b8d92c8c8658bfc4dd9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":412,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=136;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":412,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0452,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e1d0fa42fdf0d3ebe8cd","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9a3e2545d19ab4093ee68015","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"089d45c5d8094b1accaf843629ef0253ba6bf973","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":391,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=115;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":391,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-83a7e698570946f8aea9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-737a63b75a5cf82b6a0a59fc","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.2278","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"089d45c5d8094b1accaf843629ef0253ba6bf973","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":371,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=95;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":371,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2278,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b8bcef93a3158cd87361","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ca3452da8e18de92b16c43f","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_64K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.3764","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"089d45c5d8094b1accaf843629ef0253ba6bf973","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":358,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=82;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":358,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3764,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-99d409f74317c2e64e1a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-856cb23a4c100f9d9ce4d7c9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_8K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.1389","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"089d45c5d8094b1accaf843629ef0253ba6bf973","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":379,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=103;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":379,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1389,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8b5244893cea77d60960","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0bd296f88d001b9819aa32c","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6458","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a503853be9a92f6c92f00404b5d96dfa59070992","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":325,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=49;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":325,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6458,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c152dfc60dccf6778385","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c561c0f3fe8ddc8465f7f6cb","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a503853be9a92f6c92f00404b5d96dfa59070992","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":322,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=46;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":322,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-49d0330f27c12a375824","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-87d5d8ba27dfae1061ec1b8c","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a503853be9a92f6c92f00404b5d96dfa59070992","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":314,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=38;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":314,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1d45a2cee09cb6f1ee91","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-35e612f8cf086a3e40628512","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6917","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a503853be9a92f6c92f00404b5d96dfa59070992","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":313,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=37;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":313,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8f8509cc84c358602a2f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cc79d1cf35bef54be8c3dbcb","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.683","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"957585db20e136978b4255766b822dd423d71a22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":316,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=40;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":316,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e5016391d6ac145073e7","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-899241c7fe6c5765d01120db","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"957585db20e136978b4255766b822dd423d71a22","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":315,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=39;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":315,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0596a4a0ec13f6e9b32e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2e88669d9c12d19c65b8cde","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_low","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6208","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"957585db20e136978b4255766b822dd423d71a22","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":328,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=52;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":328,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6208,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-720870621b6ca8222f27","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-22810510b3411c0b1d48a4d0","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.758","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"957585db20e136978b4255766b822dd423d71a22","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":304,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=28;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":304,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b24aac1f357df457cb22","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0743eee583f9307924d6f7cc","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7583","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"957585db20e136978b4255766b822dd423d71a22","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":303,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=27;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":303,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3416733a45300d0e9189","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-92ba8317800a6f211d0185b8","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_medium","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.675","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"957585db20e136978b4255766b822dd423d71a22","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":318,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=42;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":318,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.675,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9b144d827f1185f40bce","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b3d47a44f265f31167d283e3","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7208","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34fb8a454f5235f7b6bb7b128a435f82834eb74e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":310,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=34;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":310,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7208,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cebecc549b68ab04851b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e12efa80d9fe60bd97571c16","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34fb8a454f5235f7b6bb7b128a435f82834eb74e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":327,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=51;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":327,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e6381f175b673e260bbd","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-93b26600e41a35f91ebd74d8","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7167","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34fb8a454f5235f7b6bb7b128a435f82834eb74e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":311,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=35;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":311,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0c0a028150a7bd68a197","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d3fcb1763098984d825cb5ab","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57896b99605ce6d9896c138cad765f2c75c10982","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":285,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":285,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1ab2c0929de0c6ec8cb0","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-79846a7da0fa5a448a0a2245","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57896b99605ce6d9896c138cad765f2c75c10982","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":284,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":284,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-faf551616b25b82e4948","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32d02a391188d95dbd3f5beb","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.9042","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57896b99605ce6d9896c138cad765f2c75c10982","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":279,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":279,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-38350909492ad22a9c35","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5c41c632082691755df29309","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94ca78f72a23d9e699ac522ccc2e304fe4ff8f82","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":420,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=144;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":420,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-742162b18d204155de31","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bcead6c91dd54a443f135bf1","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0694","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94ca78f72a23d9e699ac522ccc2e304fe4ff8f82","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":394,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=118;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":394,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0694,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8fc1c62a48656e5bb308","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4743fea940171b090960325","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_1K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.058300000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94ca78f72a23d9e699ac522ccc2e304fe4ff8f82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":401,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=125;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":401,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.058300000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d2c4b7c5764fa966b5fc","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-430db64063804a25963a128b","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.1361","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94ca78f72a23d9e699ac522ccc2e304fe4ff8f82","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":381,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=105;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":381,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1361,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-614dc446c341d32a2fbc","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-559c424fcb9338350174ba73","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_8K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0694","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94ca78f72a23d9e699ac522ccc2e304fe4ff8f82","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":395,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=119;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":395,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0694,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-189fd13a3e8cf263d063","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d34801ab03ea3fa4089df112","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6042","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6cd1cf0729b015d5c9269472e3f8534299af1518","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":332,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=56;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":332,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fdc4c12891de41337e96","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c21286b292a1f0a0fe8a56b8","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6cd1cf0729b015d5c9269472e3f8534299af1518","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":337,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=61;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":337,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-31","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2110e915a93d7fd200af","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2dc344ae4976feb4dbfa804","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_high","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5597222222222221","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"023083d746ae0cb1eee031fa701df427047762b0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":339,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=63;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":339,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5597222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-31","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-24bada8296fd152a8f62","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dafdf6f5778b04a369deff32","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_low","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.4597222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"023083d746ae0cb1eee031fa701df427047762b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":352,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=76;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":352,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4597222222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-31","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ffabe10b388050340588","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-074951183441b91a529752c1","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6138888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"023083d746ae0cb1eee031fa701df427047762b0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":329,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=53;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":329,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6138888888888888,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-31","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9f89be6c7647e207449c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-14bcd2e446aee776121047ac","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_none","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.020833333333333332","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"023083d746ae0cb1eee031fa701df427047762b0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":434,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=158;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":434,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.020833333333333332,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-efbdb93f385253e9c166","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e076f37f945447d1794da13f","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5972222222222221","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a38833dea8b8f110bcb4efed2eef844561b7a0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":335,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=59;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":335,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5972222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f93930956982b96830dd","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b33f164c45ceb105c90a6c9b","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a38833dea8b8f110bcb4efed2eef844561b7a0d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":338,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=62;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":338,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-05144e8014488412a775","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ab467e0edcf09ee76c04110f","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a38833dea8b8f110bcb4efed2eef844561b7a0d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":331,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=55;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":331,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-baa7705161d1dd15b408","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8dc985c426b584ce1ff122f2","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_none","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.008333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a38833dea8b8f110bcb4efed2eef844561b7a0d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":465,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=189;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":465,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.008333333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-01","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6b6236d6c35259d8fb71","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-48116e393e548f09a9c944a4","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0403","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fc49540eb5dc161a8b8de85d13e09394ef7ca92","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":418,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=142;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":418,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-35e53fe36d43ac1e9103","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-46c0062a2391ef9a6a815574","mappingCandidates":["google/gemini-2.5-pro@2025-06-05"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro_32K","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0486","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce752846b91d1e207f532dbacd06003e682c8fbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":409,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=133;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":409,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0486,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-684fe71fe8faad42234e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a7962a79f31391d8c3216f27","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7708","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1b394256fd6e8e4a0277d79f44407b3c51933a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":301,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=25;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":301,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7708,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d0c317cbe8e421079fc7","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-507c5930078d095c420f0c91","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.771","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1b394256fd6e8e4a0277d79f44407b3c51933a8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":300,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=24;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":300,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.771,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-17013883ce0042c6b400","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-625595816df4a1bf265bf7c4","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.10277777777777779","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8ba699a196ce0295f1675d69ea39197dc23f990","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":386,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=110;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":386,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.10277777777777779,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-585ce794f5f20deedcd4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9beed91931b0d7197029832","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.015277777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8ba699a196ce0295f1675d69ea39197dc23f990","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":449,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=173;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":449,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.015277777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f65719d7cefa30718ce1","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3dcdcc989c7515b59d7c4863","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.05277777777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8ba699a196ce0295f1675d69ea39197dc23f990","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":405,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=129;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":405,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.05277777777777778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-00322c869b42a43b865f","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f2ac146241074ad55bd6e620","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.008333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8ba699a196ce0295f1675d69ea39197dc23f990","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":464,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=188;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":464,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.008333333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1d50031ba97a9f27de48","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-250638ff4fe083bebb87911d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7208","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd6be18cc8b6e7c0487dde1c352b798a0127cbd2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":309,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=33;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":309,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7208,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-317ce61e185049fde75c","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c95b4a32a6dbbda19d1dccfd","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_minimal","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd6be18cc8b6e7c0487dde1c352b798a0127cbd2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":389,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=113;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":389,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5b0afdc08e82d3ac6221","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7670b2d188872692fdbe9dda","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6041666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6d512f9e05b378246b803b29b98b3a9a76ae887","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":334,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=58;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":334,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6041666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dfd445617d71cd053a22","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-95afd5c84e27c5e7ec473b9f","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.30416666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6d512f9e05b378246b803b29b98b3a9a76ae887","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":366,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=90;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":366,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.30416666666666664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fa99389c1459d7aeecb2","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b6c1f094530f9bb9d7ea3cc","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5041666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6d512f9e05b378246b803b29b98b3a9a76ae887","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":349,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=73;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":349,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5041666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-14c74e2a0c4f6859fd35","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba5ca099786e90833be4c7c1","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.02638888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6d512f9e05b378246b803b29b98b3a9a76ae887","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":427,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=151;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":427,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.02638888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3e48df77cdfe5028fc1e","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6ebd601f300b0f87f0c54913","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8458333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77a33a9fa6d33b86fda82e79d617b4a6e7710e06","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":290,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=14;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":290,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8458333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d2b562a4e29551766781","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a4e1991fa75ccc4c5b2a6b69","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5291666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77a33a9fa6d33b86fda82e79d617b4a6e7710e06","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":343,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=67;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":343,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5291666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9363352136b1bbf77d4c","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-222a754e2b55e577f7eb8153","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_medium","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77a33a9fa6d33b86fda82e79d617b4a6e7710e06","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":326,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=50;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":326,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-02-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7a66123e7d35ee5d723a","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5298f738c3120aa41ba8c520","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0486","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f4aa78e402ba34baafd8295d2c850370d68b1de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":411,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=135;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":411,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0486,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-06-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d3d0678ecdd998e50eec","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3aa27780a2e2611e54e56e9a","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.22775000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6cbcbb3431d19ada70dee3b559ed28483b29c257","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":372,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=96;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":372,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.22775000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1991234f92f4469ef052","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-74a006add7521303306e09be","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0986","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d0a8b4d9f6a236baedf19f6051db59887075ad7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":387,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=111;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":387,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0986,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-21c8769b0f3061e28b81","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b8bc03bbdde281f92fd7920f","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0194","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d0a8b4d9f6a236baedf19f6051db59887075ad7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":441,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=165;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":441,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e482aa586534d721f2f9","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4f6a3ec24d0de6f2338a320a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0194","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d0a8b4d9f6a236baedf19f6051db59887075ad7","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":440,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=164;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":440,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b290e828325d6c0509e7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-93fb295a2da8a38cdf45037f","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0749","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d0a8b4d9f6a236baedf19f6051db59887075ad7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":392,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=116;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":392,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0749,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-978568aafa58b6c1be73","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6f38bc0c2e9e20e880bccdfd","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d0a8b4d9f6a236baedf19f6051db59887075ad7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":488,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=212;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":488,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-03a861763e1b059ce127","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cacf11962a7c250daecf90e2","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b2eceba7a435b6968b797a43036ab900d332fab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":413,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=137;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":413,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-edd7ce95f7152605f9b9","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0ca85eeb4ae4eba38137c748","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0083","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b2eceba7a435b6968b797a43036ab900d332fab","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":468,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=192;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":468,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f760a3da2515d86895e5","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-19b407e1fb1c9d314dfa840d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0403","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b2eceba7a435b6968b797a43036ab900d332fab","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":416,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=140;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":416,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0403,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cd0fe99747480a88fbea","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e2456e62f7c381b3794aa7a8","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0167","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b2eceba7a435b6968b797a43036ab900d332fab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":446,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=170;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":446,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b8d302db1a2514ac6bd9","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-316d2320396a1eb5fd0ecd75","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.026100000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42d0e02d3d8684ad7bcde6b2d67fe464e9e5ae6e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":428,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=152;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":428,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.026100000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fb60a32b7c8b6302e908","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b54876a9f0e33926f8340faa","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42d0e02d3d8684ad7bcde6b2d67fe464e9e5ae6e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":489,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=213;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":489,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-927cecb06ea36c507c8b","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73bf205c4290c5d8989d5de2","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0088","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42d0e02d3d8684ad7bcde6b2d67fe464e9e5ae6e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":462,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=186;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":462,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0088,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-250c00adab498ada636b","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6d80c09544c57188156bebad","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42d0e02d3d8684ad7bcde6b2d67fe464e9e5ae6e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":491,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=215;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":491,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-99de26e90d27d9a22c95","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fbeaf4858ffb07bfebc01adc","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.1764","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea4391e138ac4e6f299f75d403f2cd1f28b39805","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":377,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=101;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":377,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1764,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1a34d524d9477087683c","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80f9b7e684bbf904fd5b83e6","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0194","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea4391e138ac4e6f299f75d403f2cd1f28b39805","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":442,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=166;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":442,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d6530e42ea0db4f6f05c","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2a0a3ca2eaeda7f4502d0b02","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0653","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea4391e138ac4e6f299f75d403f2cd1f28b39805","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":397,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=121;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":397,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0653,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-15877c1542c8f5c8b920","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a7fa6d643cf441a5e6d1f81e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0042","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea4391e138ac4e6f299f75d403f2cd1f28b39805","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":474,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=198;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":474,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6504b0db3f79d17acd26","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7fc25ec5a6b446c1024c991","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.4333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":354,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=78;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":354,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e1fbd4648c1ef7b12959","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e3f5ccb5a12db2d923e738e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0972","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":388,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=112;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":388,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0972,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6ca400732193e22db94e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6fe2de3e1a14639e6a79bd43","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.2667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":370,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=94;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":370,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3c4caf9efd5a78e63ca0","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-70e898046700010c584631f3","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0083","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":467,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=191;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":467,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-94f523fbd060a96bf8de","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88ec01be3dfcecd8d4d4bbe1","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5291","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":344,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=68;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":344,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9c2a1212dbbfc6858ae3","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-38e16992d5909881ede4b829","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5291","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":345,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=69;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":345,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ab6429d5c47a08a8b19e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32441847c6e40713074144e5","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5291","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7214f98ee631237ad862f23f4c2493de8825fe82","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":346,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=70;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":346,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5291,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a4f3e82e22d76dd71fe8","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bd78ee464d8c92e0c1233ccf","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5416","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e49811810ea491559a6986b80aced0392dc6931f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":342,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=66;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":342,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5416,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b9f71a29d7adf58d9260","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d269f45b4160ff71d9c375e8","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.3847","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e49811810ea491559a6986b80aced0392dc6931f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":357,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=81;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":357,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3847,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b0e06cf3e13a357ad662","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d39003ad4df5892e677ec7e1","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.675","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"981e574f977355b81fd374d52eda2a491b9c9deb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":317,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=41;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":317,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.675,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e96325f0d26c98321ec6","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6a9b8aa67557eea4f4c2b301","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.2917","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"981e574f977355b81fd374d52eda2a491b9c9deb","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":368,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=92;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":368,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-edf9b814d4e9c449aa6f","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ef89146f2618d4e1a7eca39","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5542","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"981e574f977355b81fd374d52eda2a491b9c9deb","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":340,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=64;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":340,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5542,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-36395ffce7b0c367a687","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86595bb0b05e946c711a757a","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7395","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"981e574f977355b81fd374d52eda2a491b9c9deb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":308,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=32;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":308,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7395,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2319c74c1c54a39d71eb","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-454e35a82a355920304cb4dc","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7395","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"981e574f977355b81fd374d52eda2a491b9c9deb","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":307,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=31;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":307,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7395,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4f0324d1577609b17388","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b8d50051c9c15365645c5bbc","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.1319","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd64f3cf5e622bc9eeac1dfa182414313ce3583b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":382,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=106;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":382,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7a8dc9c8f1e83bc0a9a5","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5db9e7bfd9f83e2559989b3e","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd64f3cf5e622bc9eeac1dfa182414313ce3583b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":460,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=184;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":460,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-69fd3a4a179e8586c3d9","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b4994dfaf7c762f755d54bee","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_medium","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd64f3cf5e622bc9eeac1dfa182414313ce3583b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":414,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=138;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":414,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-db594d673308b4b86570","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dcca6adce75c01c6f8521c67","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.189","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd64f3cf5e622bc9eeac1dfa182414313ce3583b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":373,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=97;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":373,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.189,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0ad1c9dabb9f52d1eed7","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-67b35b90c449ecef437a7ee1","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.189","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd64f3cf5e622bc9eeac1dfa182414313ce3583b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":374,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=98;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":374,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.189,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-63718fe79b7016efe8fb","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fda373a09a8a41a004946e85","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0361","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73a665f171882938e2f7949b60b5c77db96d8056","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":421,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=145;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":421,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0361,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fb10f90fe448d16fec2c","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e92287199b55b1c6ef451ed7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.015300000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73a665f171882938e2f7949b60b5c77db96d8056","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":448,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=172;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":448,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.015300000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6de9b59bf803334a0d37","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b66738ddc15ac82e625ad17","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_medium","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0194","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73a665f171882938e2f7949b60b5c77db96d8056","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":443,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=167;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":443,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0194,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bafdfc2ca3034f0220e2","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-639555bdd7bf4e20e2845df4","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0569","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73a665f171882938e2f7949b60b5c77db96d8056","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":402,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=126;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":402,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a8bc04452b6e5c775184","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a0809e9d142650447128f7c","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0569","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73a665f171882938e2f7949b60b5c77db96d8056","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":403,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=127;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":403,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cf34ab36a1ed9ccd3613","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5cbad0c25dcb8aa9efe15593","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b98083998c607dc6396c5fdc9a63497929fae95","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":296,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=20;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":296,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e957492e89e937dca97d","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-087f97615c253901f1f74a7d","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b98083998c607dc6396c5fdc9a63497929fae95","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":298,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=22;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":298,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b7e3c98ca3373e750d18","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6634c4224c0492cfe66d4593","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8458","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19ce78f230c7a857aa3ad164954f0301525c6572","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":292,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=16;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":292,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8458,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f939d1d38217e1eb2bd1","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd281010a1526e1febfc0853","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8416","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19ce78f230c7a857aa3ad164954f0301525c6572","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":293,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=17;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":293,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8416,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-58ba6a880f5b5c858938","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-99d6a35110014abea727d0c0","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8416","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19ce78f230c7a857aa3ad164954f0301525c6572","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":294,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=18;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":294,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8416,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-628d86170a79b9917404","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b99bf17e66ddefa4c7df4dc0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6596d96b226a81de5f3a0134e2122aa37f1baded","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":297,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=21;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":297,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5af147cc902418ab38a2","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08ff65141718909cc5042271","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.3333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6596d96b226a81de5f3a0134e2122aa37f1baded","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":362,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=86;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":362,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-54bea187bb240efe9cb6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-afdbccda48671a06ea937a77","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7042","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6596d96b226a81de5f3a0134e2122aa37f1baded","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":312,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=36;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":312,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-171a0b05a41c13cb6616","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-81af8866043a774d13723f5f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.85","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6596d96b226a81de5f3a0134e2122aa37f1baded","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":289,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=13;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":289,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-aa81cbbfbc8e140d2e7a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0774ba0b56b41d6fdf421ac1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.85","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6596d96b226a81de5f3a0134e2122aa37f1baded","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":288,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":288,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-eee3c298196c9407929a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cdf490e39b67653813e1de8b","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.2931","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7ff079e23ddd72fc72d5b23993092c003c80535","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":367,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=91;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":367,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0107744ed906a69a992b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d8d6aa2443f2a5dc6f338569","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.05140000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7ff079e23ddd72fc72d5b23993092c003c80535","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":406,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=130;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":406,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.05140000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-458c7aa9953a77ba6af3","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-556d2ae836e1d5480843a754","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5954","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7ff079e23ddd72fc72d5b23993092c003c80535","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":336,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=60;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":336,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5954,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-16e00021c122a70f64f4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-340f4b8142d56334e84071bb","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.0736","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7ff079e23ddd72fc72d5b23993092c003c80535","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":393,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=117;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":393,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1cc0e5a411f513c23b86","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9f245496f5b36e7c1876e186","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.4764","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7ff079e23ddd72fc72d5b23993092c003c80535","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":350,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=74;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":350,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4764,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-168110aee0567afe3b18","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1d0c8f57aa5dcadc8030eea0","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.4764","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7ff079e23ddd72fc72d5b23993092c003c80535","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":351,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=75;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":351,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4764,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-90af13e6d69512349a0c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f86a314466c739d09e1e23b3","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.8542000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a351fbbb4461c3553cebbdbdb3c6f0c78b765ddf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":287,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":287,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8542000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f099f363dd33528d3432","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c85af1de8b533c4e47ce9e01","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.425","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a351fbbb4461c3553cebbdbdb3c6f0c78b765ddf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":355,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=79;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":355,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-31d77e1c5316b929711f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80c16492719fef5a1faa88f4","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.925","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a351fbbb4461c3553cebbdbdb3c6f0c78b765ddf","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":278,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":278,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.925,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-41fe3431584b9905a29b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e281b08c8eec4426044e1662","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6708","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a351fbbb4461c3553cebbdbdb3c6f0c78b765ddf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":320,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=44;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":320,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6708,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8d6287896e05d361f3a4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b0463dd89f46f2c47c1c2988","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a351fbbb4461c3553cebbdbdb3c6f0c78b765ddf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":280,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":280,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8842785673c32ea42d3f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ab2d1c7bd81afefe902c239","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a351fbbb4461c3553cebbdbdb3c6f0c78b765ddf","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":281,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":281,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-77514f9d2d7ea27e154c","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4b8396a39876daf7c3d22c89","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6708","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de0501a94020704c707465010b83369beb778745","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":321,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=45;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":321,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6708,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e1e1728192b93f8b8473","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4a426f37c122049c079771d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.1875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de0501a94020704c707465010b83369beb778745","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":375,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=99;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":375,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b1b4160610baf673f645","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-87714bab4a8345de2639065a","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.839","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de0501a94020704c707465010b83369beb778745","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":295,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=19;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":295,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.839,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-06baa3019bb1cf56dcf9","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f3b45746ffaec465e90e3e12","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de0501a94020704c707465010b83369beb778745","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":359,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=83;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":359,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-002b43933505783bc4b7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b69004f92f99f90197722f2b","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de0501a94020704c707465010b83369beb778745","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":306,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=30;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":306,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9b5fe56c68b5dd61a469","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-04c29303869037dc4f9c6ce2","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.7417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de0501a94020704c707465010b83369beb778745","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":305,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=29;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":305,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2025-09-19","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-be8589cbc12ed8725362","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d689582f4bac5287b4905a86","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4 Fast","modelRef":"grok-4-fast","observedAt":"2025-09-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.05280000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0df44675424f650fc51661a700f7ca9b2db9c9c4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":404,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=128;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":404,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2025-09-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.05280000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-08","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-016905a0088412adb06d","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-641c96d7808ebcc0c041570f","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5264","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2836072e5f95d06ae68c52e7669e988f1ffdcfa9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":347,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=71;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":347,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-08","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6a74872aa887d5318227","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9839aa1fc8d9a30dee27532f","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_low","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.3306","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2836072e5f95d06ae68c52e7669e988f1ffdcfa9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":363,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=87;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":363,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3306,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-08","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3d55d3db699ca46bf01e","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1d04d0287c4a6c1cc2845abf","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_medium","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.5264","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2836072e5f95d06ae68c52e7669e988f1ffdcfa9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":348,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=72;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":348,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5264,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-91a4839b8ba0761408da","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f012c23378581827b2b49997","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6513888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f97a1c00c241d3afe981bf53c21a7c42decc12d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":324,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=48;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":324,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6513888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5a00d347014b55a34f19","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9b42ac8c05233718ed4b65ac","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_low","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.2763888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f97a1c00c241d3afe981bf53c21a7c42decc12d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":369,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=93;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":369,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2763888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7c54292de341217020c0","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b7566b7519ff656ea3234977","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_medium","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f97a1c00c241d3afe981bf53c21a7c42decc12d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":330,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=54;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":330,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5268b83486ffcbdbbd17","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-297ab1f168070b64deb3526e","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6708333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f97a1c00c241d3afe981bf53c21a7c42decc12d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":319,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=43;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":319,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6708333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9307a2111114577b2450","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-76613d0e03a094a49630f452","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.11810000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4c847ae90d5d94c30a941df727c267a6232e700","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":385,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=109;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":385,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.11810000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-452f8e07db5888724780","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ab394af3fec3d18390e7d1a","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e28c5aa41a9647228ad73dbf269fe1abfde0ad48","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":341,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=65;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":341,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c8394a04a68d6e9dd087","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1ab4fa9b4477e4b4a778f1f5","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.12361111111111112","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e28c5aa41a9647228ad73dbf269fe1abfde0ad48","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":384,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=108;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":384,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.12361111111111112,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_2_external","benchmarkName":"ARC-AGI-2 (Epoch external aggregation)","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"epoch-arc_agi_2_external@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fdbe701609ba02f7c548","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-21296d7ab4a821123bbe468e","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unit_unverified"],"rank":null,"rawValue":"0.6041666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e28c5aa41a9647228ad73dbf269fe1abfde0ad48","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":333,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_2_external.csv:row=57;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":333,"metadata":{"epoch_file":"arc_agi_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6041666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1af397ce52e49a13818f","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e589c92938cc1e956124a029","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2121","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fdef5bab86930247d6fa6f2095bdd14f65b19844","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":684,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=195;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":684,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2121,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e8d9a5fab77ff7dd5f16","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-39ebf715d18bab5dda0eddd7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6367","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"330938de4b287d51bc0561e1333165016b122525","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":607,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=117;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":607,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bb558bdd52ba7e793a90","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1bab30ce164348b8abd5757a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9550000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a21fb85bcabb2aaa7abfbf306d7d654a334b906b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":507,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=16;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":507,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9550000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bcdabe037288f6aacaf5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3810149a760711b7ea894842","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.905","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a21fb85bcabb2aaa7abfbf306d7d654a334b906b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":545,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=54;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":545,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7fd505477c5e609b24ba","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b7c66d5b7e242cda97717361","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.985","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a21fb85bcabb2aaa7abfbf306d7d654a334b906b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":495,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":495,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e3633d8b786b47bac113","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-96efbb27623b6ecf546984b5","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_medium","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.925","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a21fb85bcabb2aaa7abfbf306d7d654a334b906b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":530,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=39;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":530,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.925,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-70fb98f9ffde5be383f7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b84d9027a08b117781e1f12c","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_xhigh","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.985","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a21fb85bcabb2aaa7abfbf306d7d654a334b906b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":494,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":494,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-27fe2b41038b10bdc8a3","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-acef982b7100ca5a987f7f4b","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1433","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94a29b58d9bb82eb4d3f9ca5e6e8feea608aa2b0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":698,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=209;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":698,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e3b7acd7e75a8255c6b2","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-54fb54c48737804a4c90999d","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_16K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3733000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94a29b58d9bb82eb4d3f9ca5e6e8feea608aa2b0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":648,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=158;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":648,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3733000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b37c8ec11179d6f9e8fe","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b61b691e21394641ddfed2ae","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_1K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.16830000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94a29b58d9bb82eb4d3f9ca5e6e8feea608aa2b0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":690,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=201;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":690,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.16830000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-068757ae97a98105bfcc","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33cf92dd8971bf5b2710f06b","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4767","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94a29b58d9bb82eb4d3f9ca5e6e8feea608aa2b0","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":634,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=144;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":634,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e1dc57658e1dc92ea953","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c9a56dc82da84c745dc43385","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_8K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.255","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94a29b58d9bb82eb4d3f9ca5e6e8feea608aa2b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":677,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=188;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":677,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-66e8abb7748e11a7230c","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4ecc64b43df2855bc31db99d","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.225","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b2b7cd9d89b49fa43e2e65f8d31f1e68c695ff4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":680,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=191;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":680,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.225,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-20a75dd8795b6b79cea3","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-74f8f4d8faf620db630d8ea0","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3567","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b2b7cd9d89b49fa43e2e65f8d31f1e68c695ff4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":650,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=160;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":650,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-29e745cab0a97e7509e9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3885aa026de0dbe5cde2c53c","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_1K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.27","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b2b7cd9d89b49fa43e2e65f8d31f1e68c695ff4","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":673,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=184;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":673,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.27,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-995ba47c359e018a4cb8","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dea6cb329a3a13ac470d97eb","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_8K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3067","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b2b7cd9d89b49fa43e2e65f8d31f1e68c695ff4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":665,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=176;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":665,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d246917c1e23cde92cc9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-870eb2d6d76c2e165f852d06","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":645,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=155;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":645,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-670d346534aac8f43fa4","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53dee251d96b462c75a6d5ef","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.72","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":597,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=107;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":597,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.72,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d9c451557678efbd15f9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e913f1c03bb542667801ecdc","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7583","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":589,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=98;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":589,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-66b5bc8b2d154028bdc4","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9222f7f35638951287ff069","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.758","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":590,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=99;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":590,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1388084791b93721c04d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2e10c33b3163d925347672b","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_64K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":579,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=88;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":579,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ddbd8a2adda604ab9d9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-acc1052941321281ec1f7223","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_64K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":578,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=87;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":578,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-070d2d3723b1d410490b","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8a8464267ad81a7375478ab9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_8K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5867","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5eff3e1841302fcb0b8c334a3b78cc91a4b45","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":614,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=124;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":614,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5867,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b64e65b7e1d1760102bd","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eb80f723deda646fa51f81e7","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9400000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"707e02b7d90bc5a5e6e6eab02da34dd31af06c80","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":518,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=27;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":518,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9400000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1f300b7d92abdb395a86","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-beea5e1b24c2ed1ac3d00bb4","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"707e02b7d90bc5a5e6e6eab02da34dd31af06c80","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":525,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=34;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":525,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bf4e63ec524763ff076c","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8cdd1259ecc29c4c4a60a658","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"707e02b7d90bc5a5e6e6eab02da34dd31af06c80","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":535,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=44;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":535,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d01b8c7a469e0fa793a5","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-265d12139236d5d7f8f013b0","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"707e02b7d90bc5a5e6e6eab02da34dd31af06c80","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":569,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=78;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":569,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3f03006ffc0d360fc445","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20bb7638d96f5c54d65932e8","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.94","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"707e02b7d90bc5a5e6e6eab02da34dd31af06c80","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":520,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=29;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":520,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3422e045e0ec94ec1695","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6bcce2bff4eb81988f03bb3e","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.935","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3ed482869e63fea68fd060445b21eae9a241666","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":523,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=32;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":523,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.935,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6635265621e178ba8384","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-234d5b0b60c1d2a0b7b49031","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.935","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3ed482869e63fea68fd060445b21eae9a241666","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":524,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=33;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":524,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.935,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d9489debb6e2e0fecb71","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e2012a14a7189e3a923ce2f","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_low","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3ed482869e63fea68fd060445b21eae9a241666","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":540,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=49;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":540,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-942e476c602eedce53dd","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c9486ca69bc99536e46f7312","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3ed482869e63fea68fd060445b21eae9a241666","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":532,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=41;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":532,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-999c3c62a4124b7d3980","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6f1b4a693c3728e9ea531fa7","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3ed482869e63fea68fd060445b21eae9a241666","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":536,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=45;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":536,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3ca89b8b035995dd4be5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b1977b734acbb0ab2032ff16","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_medium","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3ed482869e63fea68fd060445b21eae9a241666","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":541,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=50;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":541,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-beb65e673fde091facc3","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b86081ef4df72c7080073c77","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ef34d867217f23666a8d8732b1637d316dfe96b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":533,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=42;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":533,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-68e41a71b3d29954cd65","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3e06a390e56d3cb96d45f8c","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.88","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ef34d867217f23666a8d8732b1637d316dfe96b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":550,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=59;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":550,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.88,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ebe832d90f860518e64f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba0b957ea1ba7b09ad4c0a16","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.925","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ef34d867217f23666a8d8732b1637d316dfe96b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":528,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=37;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":528,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.925,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c4662795afd388ad33f8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5967975c4f5abbbe150b2475","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.915","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ef34d867217f23666a8d8732b1637d316dfe96b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":537,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=46;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":537,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.915,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2908073f456e42df16a3","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-38ab2ce1db079da5025de074","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.975","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae102051f462cf3b2f9af1d474b373ccc89559ff","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":499,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":499,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d9c8fc811453df356112","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db1e418e6251a87bfe7e504b","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.975","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae102051f462cf3b2f9af1d474b373ccc89559ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":498,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":498,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d3c45bfcbedb1e450438","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-407fd6ef17bc4f1ed36498a1","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.975","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae102051f462cf3b2f9af1d474b373ccc89559ff","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":500,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":500,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bf8df00534f8f2438afb","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-941e9e394afdb2f8f697de45","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.255","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9788a76e50baeda04a0d789f9d6a82aad6487db","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":676,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=187;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":676,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f451685e96f2b4db6191","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6b273019b27620906ba575b4","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9788a76e50baeda04a0d789f9d6a82aad6487db","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":633,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=143;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":633,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-426324ab1ebad251b965","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-16c14a5808c2dd346d30be7c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_1K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9788a76e50baeda04a0d789f9d6a82aad6487db","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":663,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=174;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":663,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-357748570c3843038c7d","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6bc1fe895607b83484c9fb4","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6367","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9788a76e50baeda04a0d789f9d6a82aad6487db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":606,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=116;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":606,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-052a572451c97db2a6de","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-035cc1c66b5b9f57554fcf7b","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_8K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.465","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9788a76e50baeda04a0d789f9d6a82aad6487db","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":635,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=145;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":635,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ddc0f5f9b55453be834a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fadbfe5799bbbba64baefc8e","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.865","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b372725dbe1c337c690d7eb2b29e0223c7c26ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":563,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=72;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":563,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.865,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b25cac2b33c82472c70f","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-047929dd5fc0c592e7c72e86","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b372725dbe1c337c690d7eb2b29e0223c7c26ca","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":568,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=77;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":568,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d75cf746ba710c577e64","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bd18a13c67330d2b8645c3ba","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_high","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa3ab3b51522b8b000bbe34b15844fca6ad170b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":559,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=68;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":559,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cdfab4c00027f9ce23da","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3335bc7af2fcab207b5517ac","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_low","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.84","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa3ab3b51522b8b000bbe34b15844fca6ad170b8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":575,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=84;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":575,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.84,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cab973c9f4c9665bf5b5","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ca1c67c930f7f867757094c","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.89","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa3ab3b51522b8b000bbe34b15844fca6ad170b8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":549,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=58;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":549,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.89,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4f550a992e67600bcbb8","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac2c47c9331d639d0604a768","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_none","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.11833333333333335","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa3ab3b51522b8b000bbe34b15844fca6ad170b8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":706,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=217;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":706,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.11833333333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-24885c29a60d2019ab14","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bac043bc9e6196025ce6c05","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8716666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e944e3634b289de3cd966f1ee4c3c0858111bcc1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":558,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=67;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":558,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8716666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-932ebb75282afcc15e31","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-de6fdc3410d82f3f67eb192f","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.905","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e944e3634b289de3cd966f1ee4c3c0858111bcc1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":546,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=55;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":546,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-68e8516a498289b01bd5","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6fc11ebe4ba54037b5ba7222","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e944e3634b289de3cd966f1ee4c3c0858111bcc1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":547,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=56;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":547,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-174447fb859c4b9726db","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30c2b461686cb6b7bc3c252a","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_none","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e944e3634b289de3cd966f1ee4c3c0858111bcc1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":703,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=214;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":703,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.13,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f6f10d4dcb599191b229","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-60f505378d4d11b23b5bd5fb","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.57","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"913a0e724e4e769117c7046c64a3476a5255f9ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":620,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=130;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":620,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.57,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-24d4b243af15012e413a","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f087b8424109b96a101d4bca","mappingCandidates":["google/gemini-2.5-pro@2025-06-05"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro_32K","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b63e862a91dcda670c6e7969f8e277862d0fd6aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":649,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=159;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":649,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.37,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-78e9a267c4ccbc2670b8","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b00ed1c1bbdaa32322c8898","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"353f74f49ae5fcc3239f22c416135006971dcbc6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":496,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":496,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a89e8b85851b5eb04c6a","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-19d1c4ac251ea57ad24c5859","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d89aabd6aeca3021f8afba7ce4381fa93cd514d2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":626,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=136;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":626,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4263cca026c136425be4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-155d234bff6f2459cbd766f0","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d89aabd6aeca3021f8afba7ce4381fa93cd514d2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":689,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=200;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":689,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-01da08441a5b74ceba83","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-67ca3110e53a6e351685b3d6","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3233333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d89aabd6aeca3021f8afba7ce4381fa93cd514d2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":659,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=170;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":659,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3233333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc96ca1982a3a8a095d6","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ca435d5c44d6fd5387afe41b","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.075","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d89aabd6aeca3021f8afba7ce4381fa93cd514d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":710,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=221;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":710,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.075,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-79946752d6eaa8a8bf0f","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3924f7470bc98ebbf3f5cbd5","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.925","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fda2169e80c37b3deb75b7dfa700058a74d29dc4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":527,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=36;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":527,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.925,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e2db2c66efc1ecfcbff1","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2369d625516e3e6c5ab9a6d7","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_minimal","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4883","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fda2169e80c37b3deb75b7dfa700058a74d29dc4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":630,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=140;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":630,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4883,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0f51162d4eee4c9632c","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cf4d5b4adebc0a780ac5c7ca","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9116666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5896e9ac626726ce086eaa950b5f153170cee0e7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":538,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=47;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":538,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9116666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1d40f43457a01b85231a","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ff83671bf3978ab9753ecf0d","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.765","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5896e9ac626726ce086eaa950b5f153170cee0e7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":587,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=96;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":587,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.765,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d07ea158f10b98a1b75","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ada0dc3b83ab24d08cb06f2b","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8316666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5896e9ac626726ce086eaa950b5f153170cee0e7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":576,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=85;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":576,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8316666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e1c26d97c37b1a02691","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bcd87f75ce0129382e33e78f","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.34500000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5896e9ac626726ce086eaa950b5f153170cee0e7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":651,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=161;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":651,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.34500000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-31bf172fa0c7c77d2377","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-41f98e79f47571a7d8284379","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9550000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c69f3b307c1b67207e43163a1bce79a004ba658a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":508,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=17;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":508,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9550000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-437699450c10589ef8dd","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7783d2bd3c04f2d019ac9c2","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8516666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c69f3b307c1b67207e43163a1bce79a004ba658a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":572,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=81;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":572,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8516666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-56d13fbddf87d79f951b","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-394d4d958f3d59f60fc605b0","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_medium","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9116666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c69f3b307c1b67207e43163a1bce79a004ba658a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":539,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=48;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":539,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9116666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec0899c26d32545a2fe7","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-45d45b127eca02a346b75b5c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.44670000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf1421c529b05095535a4d401aa0a1ec7f2094c9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":636,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=146;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":636,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.44670000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f58baf97637e35cf3760","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0daf3857ec28f28193ffa1c9","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.77","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b600991543ff5fc7a5154b58696a183a75831b04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":584,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=93;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":584,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.77,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-872892f17aee043679c2","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c1579dd52541904e11fe5826","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6567000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc718bf81c18b4cd1ffa59fea1660cb8dfb895b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":603,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=113;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":603,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6567000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5d9516d9a358abb39da2","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-286da57b80e75b89bf268b96","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.44","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc718bf81c18b4cd1ffa59fea1660cb8dfb895b0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":638,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=148;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":638,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6bd22dcbc83e977272ed","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-06391f3c10451c32eb1a6144","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5617","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc718bf81c18b4cd1ffa59fea1660cb8dfb895b0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":622,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=132;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":622,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee7ff4e1e1d58e362803","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-529313ee5731defdc55952e8","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc718bf81c18b4cd1ffa59fea1660cb8dfb895b0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":712,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=223;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":712,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-37d3a9434745c80ba23e","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ef9efa68e5b721b7a382afce","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5433","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d922cfcd88a0362498b9ca5ad87aac7771b5b1c0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":624,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=134;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":624,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f371d8a9aae348998b8","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86eb4523081d74fb290b867d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2633","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d922cfcd88a0362498b9ca5ad87aac7771b5b1c0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":674,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=185;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":674,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2633,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dee63cc9da8a03171215","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e2f30b147db1f9159f1d7901","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3733000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d922cfcd88a0362498b9ca5ad87aac7771b5b1c0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":647,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=157;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":647,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3733000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-626d14f5d2712cb125de","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aea872d4bfdd5bc09ab490c1","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0533","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d922cfcd88a0362498b9ca5ad87aac7771b5b1c0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":718,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=229;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":718,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-86f328d74959c52b27f1","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4af2814eaf1d70d5c3135aa","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a006c25266a95c3957ec8a512669c34058dc7c14","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":691,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=202;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":691,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e592bb0ef7906ca3b9ec","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-856a7e09f415f711465aa3d2","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0404","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a006c25266a95c3957ec8a512669c34058dc7c14","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":722,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=233;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":722,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0404,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3e24ea638db6d999a389","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-70fc2dc9ac73808511170124","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2071","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a006c25266a95c3957ec8a512669c34058dc7c14","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":686,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=197;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":686,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2071,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f93c20466389b70402f5","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-44df386973e61c93c8f39add","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.015","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a006c25266a95c3957ec8a512669c34058dc7c14","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":724,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=235;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":724,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.015,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-37f963579f40aeaf687d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b450a6bb51fd15ea8ed19c7","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.728","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780d7ffb2520bc97371d1ae2f12b60eb7e3ff67c","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":595,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=105;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":595,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.728,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc3107a8223f55bd346c","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c32cf6b95a95a22aa30fae7","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7283","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780d7ffb2520bc97371d1ae2f12b60eb7e3ff67c","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":594,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=104;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":594,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7283,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-633a3629336e08121b58","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fde8716ed0017a652d220491","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.33170000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780d7ffb2520bc97371d1ae2f12b60eb7e3ff67c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":656,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=166;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":656,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.33170000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f19fbe6d324162c11be6","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-da2b9b8e2e25a0bb4ac4f398","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.577","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780d7ffb2520bc97371d1ae2f12b60eb7e3ff67c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":616,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=126;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":616,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.577,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-26da3aff1a3018c70e35","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d5a663597ac61c1816d3378","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5767","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780d7ffb2520bc97371d1ae2f12b60eb7e3ff67c","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":618,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=128;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":618,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a91384f91d9f307fb1c","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b825c512efcc710c65fb6288","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.058300000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780d7ffb2520bc97371d1ae2f12b60eb7e3ff67c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":714,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=225;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":714,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.058300000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-012d6bd57170e1e36757","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ef17150dd6e6c133462db03e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7867000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":582,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=91;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":582,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7867000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-44eb2111106a56b62141","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53867f3bb07d9e269134fe30","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5567","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":623,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=133;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":623,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1284b30467d0abd196c7","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-38ac69b2b7fc3918e6f63267","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7267","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":596,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=106;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":596,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c18f4bc273a309e5133e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3470e82e1f9a025b7c6b7ee5","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1233","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":704,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=215;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":704,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1233,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8bec2ff1788071570a95","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6595b92f4627a41fbb532c91","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1233","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":705,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=216;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":705,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1233,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77fdf0f95a17af974a93","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-65a0e7e481558817ea99e8ee","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8617","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":566,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=75;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":566,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d99e95fcfeb2164768c9","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://arcprize.org/leaderboard","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5d5006115f7d55d131a3e6d1","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.862","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":564,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=73;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":564,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://arcprize.org/leaderboard"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.862,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9963d0bacf41550fc3e6","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b6c2754064acac6176f7533","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8617","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e465a25a097e7379047cd0f3f794cf23315ef8d","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":567,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=76;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":567,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b03e00f3edc439e5046","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed88efcd97c03760ac750eb5","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8567","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f63f4a042d956003597881ad2900af9ed95898","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":570,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=79;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":570,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-11c7c2a23df0a3f12de3","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7546ee75a473428371fed1c8","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8117000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f63f4a042d956003597881ad2900af9ed95898","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":577,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=86;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":577,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8117000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e5b2a617a618ead9fa2","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9ae426d15b49c2ce135a4b7","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.905","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f63f4a042d956003597881ad2900af9ed95898","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":542,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=51;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":542,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9cc48cbd6e7cd2724c88","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5187c3c58471fe47c7e83ef9","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.905","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f63f4a042d956003597881ad2900af9ed95898","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":544,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=53;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":544,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dade313ce00a05326d94","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0e4f6c65e0a544b80c69a80e","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.905","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f63f4a042d956003597881ad2900af9ed95898","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":543,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=52;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":543,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f5074ac94f5a1c540982","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6110ac71556f31c91dba30c8","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9267000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f1c0a9f0779b0d6393941a914d575eab41bfe8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":526,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=35;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":526,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9267000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b1700a46d32ab2765dbc","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c77f042f4a970c3e588a1d2e","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6817000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f1c0a9f0779b0d6393941a914d575eab41bfe8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":600,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=110;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":600,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6817000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b2e4dfdf2808dc92d8b2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2875c7f66b4f6a5502079dc5","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8617","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f1c0a9f0779b0d6393941a914d575eab41bfe8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":565,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=74;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":565,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-606e5d52904c0a523ae2","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd211f4bd5ebd48c170e6e4c","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9367000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f1c0a9f0779b0d6393941a914d575eab41bfe8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":522,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=31;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":522,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9367000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dd68d33a515d5c6e1a29","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4274ed88a2f3a3f6209af996","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9367000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f1c0a9f0779b0d6393941a914d575eab41bfe8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":521,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=30;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":521,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9367000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-772f6189b6c188f504fe","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f74b315513513a962121d69","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0e0323152bdafcb1dabded78ffbb3930c62cd78","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":615,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=125;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":615,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.58,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9dfd4faf233cde3bac24","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d324856b83bb12194a98b8e","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0e0323152bdafcb1dabded78ffbb3930c62cd78","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":702,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=213;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":702,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.13,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-21063d2ae5f76e112182","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f714bfde6037c5753410ce06","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_medium","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4083","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0e0323152bdafcb1dabded78ffbb3930c62cd78","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":642,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=152;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":642,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cee74870a888bb1fb35a","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c4ed74b566498b852bcf42b0","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6367","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0e0323152bdafcb1dabded78ffbb3930c62cd78","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":609,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=119;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":609,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d29db90bb261e3623ad7","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-62ccfe44d6ae91e4a3510b31","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6367","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0e0323152bdafcb1dabded78ffbb3930c62cd78","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":608,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=118;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":608,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b5d8d4a902372a503d17","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8babfe8f68fa7c1a61a8e74a","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.38170000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94623f506360df6ee019123fa078584dd217cab6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":646,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=156;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":646,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.38170000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1c53c7e7bd3a3acd77c6","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98e70bd32d0eb06dc77d8cb7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94623f506360df6ee019123fa078584dd217cab6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":687,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=198;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":687,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0eb04dd85ca489b5e8fc","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2dd040e69a43bbec0f552fa","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_medium","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94623f506360df6ee019123fa078584dd217cab6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":658,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=169;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":658,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-27ebe377fc23e3de773f","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a96ea052472294c6b93002c0","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.515","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94623f506360df6ee019123fa078584dd217cab6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":629,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=139;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":629,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0d4c84c46b74e5417cff","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1d5d21f0a580b4bad8da570e","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.515","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94623f506360df6ee019123fa078584dd217cab6","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":628,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=138;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":628,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4471f8803717662b7347","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5113a12fcac9e19f04172db5","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9450000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47258153c8b46735789b2acbe5a974a6310b6ccb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":513,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=22;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":513,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9450000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d70e5bda69a45f4bdfca","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-17c26d7e427875280da19040","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9450000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47258153c8b46735789b2acbe5a974a6310b6ccb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":516,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=25;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":516,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9450000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ba3c5527cde1fc1156ba","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ee682019c6b77b2f6ba88e2b","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.965","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8a90c2b809baf9d96fce129c5fa5755f6d08d65","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":503,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":503,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e28bea32deb3e8c8f5d1","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b6af20368570e23c8676c824","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8a90c2b809baf9d96fce129c5fa5755f6d08d65","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":511,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=20;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":511,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9500000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bc4812a870e2f5e43993","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9cd651701168b9ab669a42a1","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8a90c2b809baf9d96fce129c5fa5755f6d08d65","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":510,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=19;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":510,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9500000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-62c8ae7468248dcd68a0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2171d9a0a0f9c81288a815af","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9450000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"434c5bfa85e3e6e8c5f4d872bf3940057069f6f2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":514,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=23;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":514,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9450000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a375d943e48e97636813","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-69df6dd9b88ebf4f1d7fb5bd","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7617","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"434c5bfa85e3e6e8c5f4d872bf3940057069f6f2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":588,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=97;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":588,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e08680dbe7528e6f7cdc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ecf9974a88006220c5b7b947","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9217000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"434c5bfa85e3e6e8c5f4d872bf3940057069f6f2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":531,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=40;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":531,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9217000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a94d6ee28775b3f1b306","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-81701f05a157f227ace52c5c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"434c5bfa85e3e6e8c5f4d872bf3940057069f6f2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":509,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=18;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":509,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9500000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad0df67889978735bad4","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2a7d01499ad6e28cf13d4150","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"434c5bfa85e3e6e8c5f4d872bf3940057069f6f2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":512,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=21;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":512,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9500000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-02655bad820a5928b423","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bd32d0a1a0a2d66fa4039e1","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.765","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a9ea938b4f95629c6ac1ac2c795f7162d3f928c","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":586,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=95;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":586,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.765,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-133e348ebf34cc7765c5","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-87350f2d457c80a7b1b6a2b2","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a9ea938b4f95629c6ac1ac2c795f7162d3f928c","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":653,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=163;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":653,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-74ad63c950693bc215e1","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e8fee2f8e9790ecec3a91846","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.88","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a9ea938b4f95629c6ac1ac2c795f7162d3f928c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":551,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=60;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":551,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.88,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3c413b1005d1d227e333","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c57a0a45658f8a978c05fa7c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.565","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a9ea938b4f95629c6ac1ac2c795f7162d3f928c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":621,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=131;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":621,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7490e0521f98347d7379","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1c8d003b3e1adf4a7786e87","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8767","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a9ea938b4f95629c6ac1ac2c795f7162d3f928c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":553,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=62;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":553,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-33d579a335d006f8f1a4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f9e36c7b6bc3526ea6a30c6","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8767","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a9ea938b4f95629c6ac1ac2c795f7162d3f928c","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":552,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=61;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":552,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b9b59db5c197929d81e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bc7924b418bdc6f08cd7060","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689543af78647590dabc02642201895db96355a2","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":502,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":502,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-910ba2e934fe5a90f7b3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b777498021196563d60a7fda","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.745","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689543af78647590dabc02642201895db96355a2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":593,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=103;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":593,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.745,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-26082ef7e009925bd481","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b68412abff564dfc06807d28","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.965","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689543af78647590dabc02642201895db96355a2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":504,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":504,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32294b5cdebbd075eb4c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f3c49247cbb192bd6c9dc1da","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.925","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689543af78647590dabc02642201895db96355a2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":529,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=38;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":529,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.925,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7d4b081d868c5751c510","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be6d88916841ebf1b5632acf","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.975","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689543af78647590dabc02642201895db96355a2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":501,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":501,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1caf15e9f43b6fe04aaf","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b6e3e9e302588b38a714f7be","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.975","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689543af78647590dabc02642201895db96355a2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":497,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":497,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32ffd803d85d25b61add","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a26d55b00518946bff93493a","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c819a92e311608e60b95fd84e0ba049ff4cec5d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":534,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=43;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":534,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-061e08b5ed4c402c259d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-add26800c1c1397fae47f584","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6017","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c819a92e311608e60b95fd84e0ba049ff4cec5d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":611,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=121;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":611,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b5f4b977937a3e20fc9","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d031af764efae46b59f77375","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.965","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c819a92e311608e60b95fd84e0ba049ff4cec5d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":505,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=13;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":505,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4aab5dba3241d75643e3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3db7f03f115da7901cb8b57c","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.77","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c819a92e311608e60b95fd84e0ba049ff4cec5d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":585,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=94;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":585,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.77,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-974689b92b5afd765966","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bf18d823f0ebdbacc42880cc","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9400000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c819a92e311608e60b95fd84e0ba049ff4cec5d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":519,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=28;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":519,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9400000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-09335873a669ddf4abc3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8fa0345c80983df48472c76a","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9400000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c819a92e311608e60b95fd84e0ba049ff4cec5d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":517,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=26;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":517,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9400000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2025-09-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9bc776a919c92bb717e2","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6459dc5c49cc15fb2245afd8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4 Fast","modelRef":"grok-4-fast","observedAt":"2025-09-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e0ee4cd7851a70396354d38e6e14548b6e0ac3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":631,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=141;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":631,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2025-09-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-93171f5de539cffa5a80","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-90c8314b9e28bd9b23fd5bc7","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8567","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d451fe09413ec7463b2184638136cb6e9efe7895","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":571,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=80;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":571,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e625ae257be5e1a6b3a7","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ccdb7228eba11196f916880b","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_low","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7917000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d451fe09413ec7463b2184638136cb6e9efe7895","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":581,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=90;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":581,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7917000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f8703ea6bb28d5c6cc94","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f8b14a9ff5373a4d0816ba78","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_medium","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8717","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d451fe09413ec7463b2184638136cb6e9efe7895","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":557,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=66;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":557,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8717,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4834e5e9607150cefac1","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b2db955b1b0c9ac59267b64","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd132867539f517a32ce7a82a033f7773ca31821","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":561,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=70;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":561,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-418b21341f70b76c095f","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3455892469fe891dbcf6d2bb","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_low","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7483333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd132867539f517a32ce7a82a033f7773ca31821","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":592,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=102;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":592,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7483333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7dcf6f6d4d42e8d4651b","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9d924ea17968c2726d75faec","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_medium","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd132867539f517a32ce7a82a033f7773ca31821","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":556,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=65;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":556,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8da6e837d935b14d84ba","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f8af06f7e7cb8dea847500b5","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd132867539f517a32ce7a82a033f7773ca31821","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":560,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=69;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":560,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04cb8a2ce29f1251bfca","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1babee7ca30fdbd788ef5fe7","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6533","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f711e5a2ec7223aa8c8a42f194c407ef0835c43b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":605,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=115;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":605,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3b2aff296d857f4a7926","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c902f27dd8a228df76a3527","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8666666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24ad358e8df994185ee11bb8b988254a4d4eeed3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":562,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=71;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":562,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8666666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-df21f808cdd9be473b27","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-436f0747fb4628655777bad2","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6566666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24ad358e8df994185ee11bb8b988254a4d4eeed3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":604,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=114;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":604,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6566666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-arc_agi_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-822329350aac664d807b","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-332ab300e5da6fc764e56c7f","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.9450000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24ad358e8df994185ee11bb8b988254a4d4eeed3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":515,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"arc_agi_external.csv:row=24;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":515,"metadata":{"epoch_file":"arc_agi_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.9450000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1a7fe28f7be227792514","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d719f4be0a797ec4a8216cbf","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.312","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7f9b1e05cd434c9d614531786fcf2a06c648d1b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":875,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=16;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":875,"metadata":{"epoch_file":"balrog_external.csv","notes":"https://balrogai.com/","organization":"Anthropic","release_date":"2025-10-15","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f5ae60c5c2ff0fb10d49","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-61f7acd0546142a850ebe33d","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_1K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.312","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7f9b1e05cd434c9d614531786fcf2a06c648d1b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":876,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=17;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":876,"metadata":{"epoch_file":"balrog_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-721be01c564ade420e41","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fd3e7ced772190f869d6fe92","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.435","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"65b66b29e2b29fa396753cc7fdc5f536068a26e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":866,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=7;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":866,"metadata":{"epoch_file":"balrog_external.csv","notes":"https://balrogai.com/","organization":"Anthropic","release_date":"2025-11-24","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1a16121e5518f661785d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c233a446c9dcd99a002ca4ab","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_64K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"65b66b29e2b29fa396753cc7fdc5f536068a26e5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":868,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=9;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":868,"metadata":{"epoch_file":"balrog_external.csv","notes":"https://balrogai.com/","organization":"Anthropic","release_date":"2025-11-24","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d672966b8404d2252754","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec7253b14d7248ba527a62fc","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"65b66b29e2b29fa396753cc7fdc5f536068a26e5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":869,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=10;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":869,"metadata":{"epoch_file":"balrog_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d41c0d94699e82c4242","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-abad253da134aaa7f04fa6de","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.335","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab7c0e92de5501c72ee376dff4b5ddf6353aaa79","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":871,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=12;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":871,"metadata":{"epoch_file":"balrog_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f7a20d17a7f52a9ef6e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ea058741e5d649f0caa04f2e","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.57","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e5a8948e3229491f6d5c47b0822580913a53960","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":862,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=3;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":862,"metadata":{"epoch_file":"balrog_external.csv","notes":"https://balrogai.com/","organization":"Google DeepMind","release_date":"2026-02-19","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f64cf6fb9eead23cbcfb","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5abba5960ca2cad05d096c07","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.569","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e5a8948e3229491f6d5c47b0822580913a53960","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":863,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=4;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":863,"metadata":{"epoch_file":"balrog_external.csv","notes":"https://balrogai.com/","organization":"Google DeepMind","release_date":"2026-02-19","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.89999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-balrog_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f35e56dfc50dda749f3","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://balrogai.com/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d30a30bd8249a2e468b49a3a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average progress","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.32799999999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa6934b5ec5a8732140063943d95daacce33fa22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":872,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"balrog_external.csv:row=13;column=Average progress","sourceMetricId":"Average progress","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":872,"metadata":{"epoch_file":"balrog_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://balrogai.com/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5c981eef74b9168a127b","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-758a587ef6d62a870c487bad","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3861171366594359","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f0c8ba1d260130eef41727dc3bcc398c18d479f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":985,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":985,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3861171366594359,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-74b628c6adf185f8e750","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d3480cc1435db9a552f8544e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f596edfec895a3a9a3564f7ce0e79263b8798107","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":998,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=15;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":998,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fda073daf46491604661","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-37d74da669f6cb7aacbf0999","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.245119305856833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90887e581148969f7c16276934adb0b2ee4d89ef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":993,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":993,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.245119305856833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-786a0038d1e44c5d73ef","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cf772f4e3a44776526a002ae","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.14533622559652917","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8f3e47142f040b2d353505381259d349b07dbb17","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":995,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":995,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.14533622559652917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e4502b8c796616df26b8","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-092033e01fa8df72bf1b8b26","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 5.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_unknown","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.30368763557483736","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28efc593a8899970b020bf21c0ce7b3d67fcbba0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1003,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=20;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1003,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.30368763557483736,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b3e673c8d0acf753b990","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e17ce06f87f1859f7423f26a","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.06724511930585665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70699002d945f7c7a5f5123c8899ff0e96e268cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":996,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=13;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":996,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.06724511930585665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-de5a262a37999152fddc","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-07f52eeffb8c60339b595faa","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.24945770065075923","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f97b643d378a979a1487f845c0c29057df7ebba9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1007,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=24;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1007,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.24945770065075923,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-33ec8a9bea21273e93ea","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b69a56975eb670d9ad158b8","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2646420824295011","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"510b6700970c8c93175e02139bdc098741551e43","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":992,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":992,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2646420824295011,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75d3b539bbd329277b26","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f4f3e1e8a4963135700f916e","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.336225596529284","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f4921193dd89c77abcab45e907384dbe462c907","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":987,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":987,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.336225596529284,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b62214b972a18a02f7de","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-861e4192c89e3e20610adf9b","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_unknown","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.31236442516268986","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"568dff1c8453f7c21126bcf0da3570d876796317","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1002,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=19;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1002,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.31236442516268986,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5621a428a3b4f34e1af0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e40723742a8954f1a33f6760","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.27114967462039047","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e574d922c1304f96e33961c7be72e729052f5b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":991,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":991,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.27114967462039047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5b082fc3c3c6b85959fe","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7c44b2d16a956f5e9a593a5","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.36225596529284154","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"256ac9e84d030b5e8e9bf5081912db964f68d5e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":986,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":986,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.36225596529284154,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0cae2eb9a27ae2ac025c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b5040e71b0638941f8ecf4b8","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.22559652928416485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a7d0230245c1b28b04bcec13e7f405ce30dab1d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":994,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":994,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.22559652928416485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d5f6222c279c791383e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-257fa9d4ac3e30b311b91136","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.336225596529284","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7a50032dde640aa1b8f203c666a07425b1082b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":988,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":988,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.336225596529284,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-82dbb21f5cf41ed50b29","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e552ecf421e02294605dd316","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.30802603036876364","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d79971d54f6d5236f1822f5d6b55388e4968003","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":989,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":989,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.30802603036876364,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ff0166dee72b9fb61ce9","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ce9481f0e958bc6e35f3289e","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2733188720173536","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41be45ed68d5c28c1e8754919d8283731891dded","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":990,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":990,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2733188720173536,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-308ce3af0a77a4a33a50","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bb2b014e4555fa48e666901","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.33188720173535774","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9fd840851305a93b857162f0deb46c79a7659cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1006,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=23;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1006,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.33188720173535774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-418120207b1c5010666f","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f1bce64825a25ea615ee8cf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.03904555314533626","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60e2972cb325d5a07a5f45a9fd0ea4b7dfb9db7d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":997,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=14;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":997,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.03904555314533626,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-blueprint_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-146c3c2d67cf237dfedf","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4cc889d94545205098adab1","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.29501084598698485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22eb316eb15c6600bd63bd7492173ee71597f4f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1004,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"blueprint_bench_2_external.csv:row=21;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1004,"metadata":{"epoch_file":"blueprint_bench_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.29501084598698485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7a0e6eb8076c4929d24d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-55ebce056f60240ad7395b1a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"020fd37a1d63413f3d631c67a0b1194923c8943b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1145,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=3;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1145,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.13,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e917267e763904c9d080","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-e8341041bed00e2e16a36d2f","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efa24e1b7238370e64adeac90b776e2427c8ce1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1149,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=7;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1149,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-42589837a1edcd8c65ea","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Vendor agent SDK","harnessId":"Vendor agent SDK","httpStatus":200,"id":"pub-38652d412e7a3f501848776a","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Vendor agent SDK","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.137","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efa24e1b7238370e64adeac90b776e2427c8ce1d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1148,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=6;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1148,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=vendor agent sdk"],"subjectType":"system","unit":"fraction","value":0.137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e7c6ccdd9da2f4eabf9c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-82f64451d1ed293a97b18232","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efa24e1b7238370e64adeac90b776e2427c8ce1d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1144,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=2;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1144,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.13,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cdfed32ed16b66ddeef0","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-aa726b5890057494f312b12a","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_xhigh","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.118","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5526a578a52aaf25948ec79c67771e2b8116ba0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1152,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=10;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1152,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f3cc6ffacd9131eb84d5","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-d523de5cfbc7a718eecba119","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.154","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"61cb15166158ffd7452f0c23d62e0c5f756f3fda","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1151,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=9;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1151,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.154,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-01c903fa4e95aedaaaf6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-f306077ce0ea72e00e8f41fb","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.143","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ea5f96ce617fd3bb00608aa9412594ee2d99e1e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1150,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=8;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1150,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a2eaab369c87ae149566","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Vendor agent SDK","harnessId":"Vendor agent SDK","httpStatus":200,"id":"pub-bd690d0d28ad1ad42e00e672","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Vendor agent SDK","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.134","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ea5f96ce617fd3bb00608aa9412594ee2d99e1e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1146,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=4;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1146,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=vendor agent sdk"],"subjectType":"system","unit":"fraction","value":0.134,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3846ef726cd79cfde997","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-bdf8b7aa1fcf5ecf01b400c4","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.135","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9776139e4ec80024c5a20112d202ff6707c342b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1147,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=5;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1147,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.135,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-btf3_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0a2eff61b83c1d168a93","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"FutureSearch agent","harnessId":"FutureSearch agent","httpStatus":200,"id":"pub-6c4da5d1e5f3a69a5dd5320d","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pooled score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"FutureSearch agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.137","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9776139e4ec80024c5a20112d202ff6707c342b6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1153,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"btf3_external.csv:row=11;column=Pooled score","sourceMetricId":"Pooled score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1153,"metadata":{"epoch_file":"btf3_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=futuresearch agent"],"subjectType":"system","unit":"fraction","value":0.137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-98c25cf2f77e729f49cc","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-038ef47437cf98c11b99748d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55965d9d23d23b381acfb0381b530609f6676931","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1160,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1160,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.000000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8202907f05823668978a","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-012fc5e65fcf647dcc8833a0","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.41","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"063f8dc1abca7fcaff943699d2b3c37990afa652","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1233,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=81;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1233,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4b97c9a0592125a4ad28","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d94276a89b90000e0cd19a83","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.28","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"063f8dc1abca7fcaff943699d2b3c37990afa652","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1232,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=80;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1232,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4a4988d251d46f39de50","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe5bec1d97eabe441c39785a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.41","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"063f8dc1abca7fcaff943699d2b3c37990afa652","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1279,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=127;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1279,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-79af7f200935fc38742f","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-273d4414f198e3a071cfc5d3","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef1658d2f02ceb3c979d978653f8406b39f77a9d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1243,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=91;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1243,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c482e44f07c7a23e9d3d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3de9ea73e57e6634d9d46f1b","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3896adb691aedf81a9e31893e15b9343dfc1e6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1231,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=79;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1231,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9926313eee2d74c6d707","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2d3526f9b71fed1f32ad3df","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b3896adb691aedf81a9e31893e15b9343dfc1e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1314,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=162;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1314,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c9b111e1f60074723c7d","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7917ba7612edfc4a880fe5f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e752caeb695c5b08a05d0256cb83fae365287f65","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1290,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=138;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1290,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a9c1bd825ab2432cace2","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-26746e3d6dab1441ab9f57db","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e752caeb695c5b08a05d0256cb83fae365287f65","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1295,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=143;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1295,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c4608bf8782c2b8c2698","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bbd202b92ec72f3e04b2004","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_64K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e752caeb695c5b08a05d0256cb83fae365287f65","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1294,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=142;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1294,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f789717471f5321faf99","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e43e2a0c0db2dc13bad94eac","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e752caeb695c5b08a05d0256cb83fae365287f65","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1214,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=62;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1214,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.000000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-00614e70648ca0753c9b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-56182c935b3ff9a4c85fb1a8","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_low","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a3b52fcfb0e120fe0fc3898763c09ac66990685","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1261,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=109;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1261,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b0a422c624f1b7cdb355","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-619d4e18066ee2c0b0e2bae4","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a3b52fcfb0e120fe0fc3898763c09ac66990685","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1230,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=78;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1230,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-194e8a5e6b0345413508","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b8634a19537dd471df5460e","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a3b52fcfb0e120fe0fc3898763c09ac66990685","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1285,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=133;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1285,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-49362e26bd91c2915f8a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6398eb60d1912b520a4e077","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf73fad83d62d1a7c422c630c18aa7ce448dd54a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1229,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=77;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1229,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.999999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-71a07a7024f4722c7052","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-de6a1a90c76bf96a969988b8","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.34","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf73fad83d62d1a7c422c630c18aa7ce448dd54a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1280,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=128;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1280,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-80dd631902aa7b6c52f0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8370278f939efc16521e6198","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf73fad83d62d1a7c422c630c18aa7ce448dd54a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1228,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=76;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1228,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cd4846e386619345a6ea","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-62d94ddb70e817ab22c38604","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5b47a86e81c5b4644a467009ce14efd53b9609b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1227,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=75;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1227,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a19ff54061768850aed3","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3324addc998ad9eb86234dbe","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5b47a86e81c5b4644a467009ce14efd53b9609b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1226,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=74;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1226,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c980398b1f4ad233ca9b","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4fe5acb3f3f99e836fb3e592","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.42","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5b47a86e81c5b4644a467009ce14efd53b9609b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1240,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=88;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1240,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4b9ab05be640797ebec3","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a95fe7cdbcc3e3e90240d570","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"825f1b4d62feeef055c70930fb0315defb5e8824","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1224,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=72;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1224,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-44fbcc7bb0377b5a1e4c","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2a8ce9e61a1639dca94f3bd","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"825f1b4d62feeef055c70930fb0315defb5e8824","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1309,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=157;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1309,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e00d9c58ccda28c1c276","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0b1d5b80f8c5a682687b0da","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_32K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b197c165ea94ee85f7e4b4746bc60305e8bebc0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1291,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=139;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1291,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-20798326e9cc968dde4a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7938d65003a5c6220a7ddb85","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b197c165ea94ee85f7e4b4746bc60305e8bebc0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1263,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=111;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1263,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fcb5cb336dea96887611","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0a9f27849af087e93a8536b3","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b197c165ea94ee85f7e4b4746bc60305e8bebc0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1225,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=73;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1225,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-84f2896a941556713556","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1856ce8e0c30e4a3b6474808","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b197c165ea94ee85f7e4b4746bc60305e8bebc0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1266,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=114;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1266,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-013cc1d7dfd28da9a8f9","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3875977ae2a6796636145ee2","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84404411a70f484142466aa7e4c410f3bbaaa9b6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1275,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=123;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1275,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-85d7636ef3bb940c658b","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d5ebdafca01b76b9060f7fdf","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84404411a70f484142466aa7e4c410f3bbaaa9b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1274,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=122;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1274,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d3ded964ea9d882d47bd","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e472029872c07b6d676b48d1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V3.2","modelRef":"deepseek-chat","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.01","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"849fa74290bb0d4eea0b0a12cfc7221ac441915e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1244,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=92;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1244,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9360dc197c34e0e3ef56","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-51a404e9c6fb7e191c5dd8ed","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d15bb30aad493478a299d45e6a80fa545e7326be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1238,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=86;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1238,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-736fe07060b606c44afb","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-739377b74349e3cf9260585f","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dfab4a4abe03de7ddd12f080b5dd6552992b7987","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1155,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1155,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-345fcc77686ad2a57df6","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4c8afb6e4203b5df65703bb","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"408bdf8a787910ec21caef1dab888065429b1252","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1213,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=61;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1213,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2133ac34570f99133702","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6c2c894f5a8077059360909","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"408bdf8a787910ec21caef1dab888065429b1252","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1278,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=126;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1278,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2315225bc0b7d6ec9974","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-160b76a8645192d7a11582b5","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"408bdf8a787910ec21caef1dab888065429b1252","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1212,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=60;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1212,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee906b3fbe5ec8f8c9bb","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-61f944c64d98c7fca0c00f6c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f13337edc96f476c5c9015db1fa2ff0a3ceff1ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1298,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=146;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1298,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bb7e1f4cacd4ce9584ff","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5fba340f8eded73510d0db12","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b6142ef457277c23d9033c05f13d5b16b6e16dde","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1315,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=163;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1315,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-53dddb712d31ab934472","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a959816d0bd34bd235e4caa0","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a4359d858a60976301318dd98208b3cc2f48ce3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1292,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=140;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1292,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.00000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-240b829d545f6f0b23b9","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1fd973a83793aeac96fa04d4","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a4359d858a60976301318dd98208b3cc2f48ce3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1219,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=67;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1219,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c9a9e75f466986b34789","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b06b5017adbbe611069117eb","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e16ca5a949257c9d2b1a4d1e49389f1c7d87977f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1217,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=65;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1217,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a6363776b18bc8b4eb4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-878288b19809c4055a71ca20","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e16ca5a949257c9d2b1a4d1e49389f1c7d87977f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1216,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=64;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1216,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b2839550a8a36786db37","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3493f8118b9e5b1be3dba352","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e16ca5a949257c9d2b1a4d1e49389f1c7d87977f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1215,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=63;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1215,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0cf337f5728985c95a88","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8b855b61acf0f9b564a8d3cd","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7deea7ea54f71eba2b03185956bbe3baa8b40d4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1281,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=129;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1281,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5c18f50f1e03988c6e23","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c1a03bbdb8c9e3f08de0a562","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_low","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7deea7ea54f71eba2b03185956bbe3baa8b40d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1218,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=66;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1218,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a161982cc37649fc70ef","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73d9bf60a3dcce51423d7784","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_minimal","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7deea7ea54f71eba2b03185956bbe3baa8b40d4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1259,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=107;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1259,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-609a07a67a975f0bcc31","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-84d31b64fe0db4172fd4f2e0","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bcb8081f6364ff37453065172bb6fa7765da20c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1239,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=87;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1239,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-55d9243f3fd99e2eb643","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e4883bf4d141da38abcbc05","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bcb8081f6364ff37453065172bb6fa7765da20c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1194,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1194,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-685160742d81bba104b4","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e7897a543b08669324af4218","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bcb8081f6364ff37453065172bb6fa7765da20c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1193,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1193,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b2b8c236094cb33e031b","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e669176b91a1b36b0c1b2d4c","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4be905aaf138b251ace104bed041fbfc9a71ea9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1156,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1156,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f3da3dd46bc0515b42a9","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e5d0c28a0a0dfe83c5dffc35","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2c71499eebd4bded50a45c3de46dbe62994da32c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1297,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=145;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1297,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-28cf9e7df2ec025c8688","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6a6c698ae833a36eecbc0405","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"956a2afa409157009776fca8e040eb7c2676251d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1293,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=141;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1293,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a15ffbaaa3e4f4b36105","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7a844bfa7f52e8c99bf0127a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81115bc7f1ff894973e9b4dd97af2cea432218fb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1162,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1162,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-592674c29252a02da7b1","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2f55ee0845fcfcb446267b0f","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca4cc9e9c5cc7026bb7f6aa3bdcbc567491730b1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1276,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=124;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1276,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-63f60b101ad8374cec5d","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f512d9b2b10055c203b126de","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca4cc9e9c5cc7026bb7f6aa3bdcbc567491730b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1161,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1161,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7bbfeb8fcd8b17be236d","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-007a624638b53d08afbaca5e","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.3","modelRef":"glm-5.3_max","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71d161a54b9548b4e06184c293cba500f1f0cb93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1154,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1154,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-279354b5c5ca498c88f4","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e2d44feee76e8b9e24907601","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c333dd13528e4f39201e773bce91979f073253e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1312,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=160;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1312,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a289f174d8acb1ae882","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-62804e7c1d1ffb2c4f449890","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c333dd13528e4f39201e773bce91979f073253e6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1257,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=105;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1257,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75d3c82e78089b26f80b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-015226ef900922dfaf06cf31","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c333dd13528e4f39201e773bce91979f073253e6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1256,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=104;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1256,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.999999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-90fd52feb64f9a001afa","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b40cba42260f7fe3e1e45c3f","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c333dd13528e4f39201e773bce91979f073253e6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1254,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=102;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1254,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0299c6b0173ca35a25e8","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9f1bfa0185d1ffee7a54861f","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"590e4eda6474ffe15b55dcb919d01cce80cb6c8e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1202,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=50;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1202,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e71f444719bc141d575","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-754007ceb7c519bb9a03e7a7","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"590e4eda6474ffe15b55dcb919d01cce80cb6c8e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1203,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1203,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f1ceccfec5dbfb863aa","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cef9b7f635c61a8aba82a82d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"590e4eda6474ffe15b55dcb919d01cce80cb6c8e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1201,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=49;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1201,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-17a0f606f9c06c4cb3d3","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b9b0cd80a89bf356c10a0689","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.27","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8cc5f89c345cd3de454ddc751c98b3b39de6816","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1207,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=55;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1207,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cecbc1984cecd1b8c46b","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4a3fdf2f940cd82005c837a5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8cc5f89c345cd3de454ddc751c98b3b39de6816","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1262,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=110;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1262,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-61f88df669bad7fd981d","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2b7757f20ed9be4863eb7dfe","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.01","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8cc5f89c345cd3de454ddc751c98b3b39de6816","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1264,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=112;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1264,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d09675ac2aee5cb5121f","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9019ff7263d99e405880fe2a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.32","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eecb62ef6bcc4d93d748cc1e65153a3673cf4d5a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1311,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=159;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1311,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-43a200028665d44e7c37","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-22f4ad1638bb97f0cafa9231","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eecb62ef6bcc4d93d748cc1e65153a3673cf4d5a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1191,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=39;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1191,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.000000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ea356823140e644769bf","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a11b58d8218e5f71dc0b8aac","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eecb62ef6bcc4d93d748cc1e65153a3673cf4d5a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1179,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=27;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1179,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0398a9d5a148e99eda70","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3dad1cb2980de36c93aeec9c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41b17ca884a0cf2bfe62fccd82004e403bd244fa","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1304,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=152;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1304,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7004008467c30f22ad9a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-35dc567736ce613ea58ea4bb","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41b17ca884a0cf2bfe62fccd82004e403bd244fa","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1305,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=153;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1305,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f3ab1fbea97b9fa159e8","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-292f5cd8d48cdc8066c0e079","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41b17ca884a0cf2bfe62fccd82004e403bd244fa","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1303,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=151;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1303,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-89a0d78dfed0b6ad3342","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-972ae6328adf3a105f9b867f","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41b17ca884a0cf2bfe62fccd82004e403bd244fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1265,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=113;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1265,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad196695f301d6815c63","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-156eb016f1a1f8ba835ac7ec","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41b17ca884a0cf2bfe62fccd82004e403bd244fa","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1301,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=149;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1301,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e5d1e5b5d5fb6832f67d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e41bc566f8ec9c55a2169bb","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb3023085cf68d91bae42c26963979e162413ab2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1252,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=100;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1252,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f5881a15eb68aa7bc011","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-69fa49bc70d9e933bba8e992","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb3023085cf68d91bae42c26963979e162413ab2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1253,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=101;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1253,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-db8321e993afe3be5195","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d9eb62ee83004c9fcce37870","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb3023085cf68d91bae42c26963979e162413ab2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1250,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=98;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1250,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bc7e3676a2be736a414e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d38a947c1001526c013812b7","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb3023085cf68d91bae42c26963979e162413ab2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1251,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=99;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1251,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c1657df30b6b6fed006a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b1fb2b8c2a28a5803853ea48","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.44","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb3023085cf68d91bae42c26963979e162413ab2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1289,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=137;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1289,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b2d4c954b8f4c7d02e16","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-52dc8cee916905a208fff0b0","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8743fbb3ea8f59a00241b5a646bfee0cbed13a7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1286,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=134;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1286,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1676569fbeb68b65e683","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba0ccd40194a00a97226c7f4","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8743fbb3ea8f59a00241b5a646bfee0cbed13a7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1173,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=21;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1173,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0e66918cfeed2037ed4a","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0dbd64c73b69e95c87fbea37","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8743fbb3ea8f59a00241b5a646bfee0cbed13a7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1166,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=14;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1166,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8366601c6d526fc9921f","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b86f789a60d1381c3faa3bb7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0d96d6f0da6d9feea82dbd53467e0b9ec626236","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1287,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=135;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1287,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b7b07aa77a9802fbc9ef","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6ea57b7bf868e516445b26f5","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0d96d6f0da6d9feea82dbd53467e0b9ec626236","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1210,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=58;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1210,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1fb778df57d31a7de914","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df013abc0b830be0fc1f5203","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0d96d6f0da6d9feea82dbd53467e0b9ec626236","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1208,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=56;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1208,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bc25747aae0f65e6bb7f","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e2ab15ce31112bcfc8aef5e6","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e44b1b3ecb3a0b3954f4b981bbe129071fb4e26a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1288,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=136;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1288,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.599999999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-221d841be326763a325b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6c80d906448ca7586a661cf8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f8f881a8076a2af6969111f91f6064a84f4ea22","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1237,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=85;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1237,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d37d695b8b2386627d55","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-864d7fb20c43b3a89fc5c26b","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f8f881a8076a2af6969111f91f6064a84f4ea22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1176,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=24;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1176,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2cfaa49cdc97d0b0acc2","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-035053f23e048361bb0d93a9","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f8f881a8076a2af6969111f91f6064a84f4ea22","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1168,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=16;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1168,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cb6c38e6c17bb49d1b29","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9b178a251ed2864b759fd163","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed9d542333c641e1264c9fe67c6f8b88db37e4f5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1198,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=46;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1198,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6580b72cf72f5518b8cd","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e52d4eace77f6469c53d6033","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed9d542333c641e1264c9fe67c6f8b88db37e4f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1271,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=119;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1271,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7126f8873469e0b734fc","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-84280ae74f22bfaab29dbade","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed9d542333c641e1264c9fe67c6f8b88db37e4f5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1183,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1183,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f229194af61cad97f90","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1216d66f44090ca4ddfa880c","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.27","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6d923e315c1938043502be4705b771c90cfdb2e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1174,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=22;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1174,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ed35d298425d2aa56697","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9f59356d9396c0738bfdfa00","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6d923e315c1938043502be4705b771c90cfdb2e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1272,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=120;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1272,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.00000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-58cae37db108c5ed78b4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-426ea876e1d0b1b4de802cf3","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6d923e315c1938043502be4705b771c90cfdb2e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1167,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1167,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b6f03aad961b7245920","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-617e1bb82062f6c832c5e36e","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_promax","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.64","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6d923e315c1938043502be4705b771c90cfdb2e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1269,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=117;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1269,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":64.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8b0f175ef1fefb6cac41","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4eedcf3959632f234bed6ecf","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f6ead0890374ab1054ab17cac9c1c08bb0dae64","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1181,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=29;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1181,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d187f76cbe51cd57a030","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f5131329ba6c1f34bd5fe140","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.54","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f6ead0890374ab1054ab17cac9c1c08bb0dae64","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1270,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=118;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1270,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ef00d953cb29a7911649","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6217ece18b4de3c7abf4ded9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f6ead0890374ab1054ab17cac9c1c08bb0dae64","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1172,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1172,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1fdee82815a2be7237a7","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e35565da6b8b2cc9a7d42be4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bfa3d668bba5449f19eaa32fd6e2f309b78cf8f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1273,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=121;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1273,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-63e3bc7775f9b3b8f7e3","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-97b8f011b3fe97e8cfaa789b","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7fbac3534b2c3a32276f744227c98cd6dcb2eb8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1159,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1159,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3b9ca0ec8cdb07b9fd63","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c529cd6b85a779c433971820","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7fbac3534b2c3a32276f744227c98cd6dcb2eb8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1158,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1158,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8a5d7f4498f5d4e642d2","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-285dc1507c9181c1da720657","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a5314a2e3fc23e53d6bbc1abced204f4e8188228","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1282,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=130;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1282,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-657614225740da42e49a","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d44ea23167d6bce5bf85290e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f38d8e613d7009bf45e36e0caa41d0d360b10bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1180,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1180,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b9c23ad477ff39325b2d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-804c032cc29e3ae56b72ecc4","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2212ab9dceb1368438ad5b9536182166d3d201b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1170,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1170,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-79ffc7554f74d424f197","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20cf1f3bf60009f1656e670e","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2212ab9dceb1368438ad5b9536182166d3d201b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1169,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1169,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0ad3de15595776d5d3fa","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d377ae85b517108c0f05247c","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.39","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2212ab9dceb1368438ad5b9536182166d3d201b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1242,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=90;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1242,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f0a2fe14fca0659917a6","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3061e08f3c4600f6aef9bb44","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f10e5b332c84ab890780bf134a309d61d8c84ede","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1189,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=37;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1189,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-54f90da71c8d491fe590","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1cfad99a570bd1845d546dec","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b_none","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f10e5b332c84ab890780bf134a309d61d8c84ede","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1190,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=38;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1190,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eee4274508abca3ae2da","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1037ba670fe4ed7d901f217f","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"854199e6e18bc067e71f2775dff64f2e3cdf83a2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1209,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=57;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1209,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-44fd6907e698c650a0bb","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2758380a83b6389098371950","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"617df65a9f73eaf816b47b2b5c624ebdee153e37","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1192,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=40;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1192,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ea165f0656486e887d9a","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e8f764d2a34a2086f49bd95","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026-04-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d3da83618c18a66ca88a5ae616c965870341530","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1187,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1187,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a3ba6e9e4b68ecef9ad1","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eca80e11a36d86de141c31bb","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b_none","observedAt":"2026-04-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.09","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d3da83618c18a66ca88a5ae616c965870341530","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1188,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1188,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a95cc62c2bdbda7f0182","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5aaca41a32519cfbafdfba8f","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cca5e95f1f476b5f572df6d694dca7810c1f17a4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1195,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=43;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1195,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a7ca7e599c1444686a72","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73b82998ceb9d5374d2f507a","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b_none","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cca5e95f1f476b5f572df6d694dca7810c1f17a4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1196,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=44;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1196,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-56c3d426a5645861f5c3","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-987038b6d3129780a60d1a63","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"808cf30e26fa993acae3e6e437df492718fe8a28","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1197,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=45;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1197,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e67af1d56f331b549848","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2d1c5e4bfac2957ba947d664","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cf414104689ba240f515905742d293d6fa09a21","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1175,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=23;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1175,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6a9e118dab08acc2d4ff","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0212f15a7ade3ccb1fb828c9","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8780ed67a43583dfbf76cb15324ca6d90eeb081f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1184,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1184,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0ca22a5230890bef5f54","canonicalModelId":"qwen/qwen3.7-flash@2026-07-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a774e3fcab957c61cc7d59e5","mappingCandidates":["qwen/qwen3.7-flash@2026-07-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Flash","modelRef":"qwen3.7-flash","observedAt":"2026-07-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"279ce1b995032bb24987d64447da009c70f4f2c2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1204,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=52;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1204,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-07-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c552e58fd2c6a009c840","canonicalModelId":"qwen/qwen3.7-flash@2026-07-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9060063e875f3d4e23bfe72b","mappingCandidates":["qwen/qwen3.7-flash@2026-07-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Flash","modelRef":"qwen3.7-flash_none","observedAt":"2026-07-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.09","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"279ce1b995032bb24987d64447da009c70f4f2c2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1205,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=53;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1205,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-07-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8b96c26873a4abd885b5","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7fb8f467104f24ee5dee2a91","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"187859910b9f51ba443b1c9abcb5f7f7d845bbbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1177,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1177,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3042426d93a0528d752a","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4d254c603c87a2ecf32ebe74","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb191e90d68f9ce07950809fb43e69703202b22d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1199,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=47;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1199,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b94cd7e8ded17a296b93","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3be60c1cdab6534d2e7e571e","mappingCandidates":["qwen/qwen3.7-plus@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.7 Plus.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.7 Plus.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus_none","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.09","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb191e90d68f9ce07950809fb43e69703202b22d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1200,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=48;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1200,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-chess_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-688c3f03ebc84cf899c8","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2230dab74d046a7fe2505883","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"507924c9cce48c31286bbd03f167243b6582c3cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1236,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"chess_puzzles.csv:row=84;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1236,"metadata":{"epoch_file":"chess_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.999999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-515453c5af9e48ee2715","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-10594b9a39db5759f09049fa","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.132","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ccd0fbf2b6925aaccb840691e2e31fccfa05734","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1336,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=21;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1336,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.132,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-abc350214e6e6293939a","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-51d2513503aada81fa1c8e3c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.114","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"347f6d010c96ac18a55f5c6272824b7e7831273c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1339,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=24;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1339,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.114,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0a073f2a93c4e6d903b3","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-85aad312b561e554187e6711","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.211","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9992b195dee908d51c5d451de3097067c024c061","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1321,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=6;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1321,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.211,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6401f8438d0f04093582","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-99d5eed13b0c6f58961db61b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.207","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9dc703fd3ad685935ffe47ad1493b9ae0896b937","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1323,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=8;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1323,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.207,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-354e828a35b812ff7ea2","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4283888d4a3869944fce6114","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.12400000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a1191da37ca06ba6758cc8f0ac142858636ed779","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1337,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=22;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1337,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.12400000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-36a4c1266f77a5a0480f","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ebe35acf5ebd2b8a479fd1b6","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.20800000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37110a2cee2c2f03f01bb5c2c7a61fe9274c1c04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1322,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=7;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1322,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.20800000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9dabe9efea5bcff9eacb","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e208c8e36b99e5d14d4a4fb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.187","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88b8652d35b4377041ceb2fb4c0ee2efed22de3f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1327,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=12;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1327,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.187,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7a6275ab603e1faf1883","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-61cb9f5f4a5a8a35452eedb9","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.237","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68071495ac30e5b2c4e857ad0f320b1002d2bcb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1318,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=3;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1318,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.237,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3213bb193a29ec550d63","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-209c7b0972292e70b64258be","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.211","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68071495ac30e5b2c4e857ad0f320b1002d2bcb3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1320,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=5;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1320,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.211,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7fec5db6be314a1237f9","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20e3cdb623a551c7647f1b57","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.181","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3fece8a6427391895f374f0ad8b355e39fd3bffc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1329,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=14;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1329,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.181,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-76580d93d1690503980b","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df5fbd93b93d62878ae9b0e1","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.182","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3fece8a6427391895f374f0ad8b355e39fd3bffc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1328,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=13;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1328,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d7056a16a5cc425c679","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-81cba128cc6553d868cfb5e4","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.279","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ff782a644442b43752ba82b44810a72be8dff03","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1317,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=2;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1317,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.279,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c87be4538f5754e3701c","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fa57392fba8d817011ef883c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.193","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"483d094df93ddd55c4134b22454f67aaa26a7790","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1326,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=11;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1326,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.193,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-97d3fcebbffa639605df","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd162e39208984607a217d0d","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_unverified"],"rank":null,"rawValue":"0.157","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f46d86f0749f3779cc0371434bb7621b34a9035","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1334,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=19;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1334,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-92d08fee31cae064354e","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-709cfcae239b6533650a426e","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.198","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8260e3cbc59d5b8892313975192e58b60f8dee2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1325,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=10;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1325,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.198,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-256d560e63204d0dfa5a","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a7ee9b599328570e76cf22cf","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.203","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"320bc8005440ffd550dbf251aea0d8b82afa0a34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1324,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=9;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1324,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.203,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5e27e6591a52940c0469","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-741f8fea80bad2a285278108","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GLM-4.7","modelRef":"zai-org/glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.159","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c40424f3b8960f34e9e501a30ccb1a4f4f833811","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1332,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_external.csv:row=17;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1332,"metadata":{"epoch_file":"cl_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.159,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b3099fdfe000102f982a","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b2343ed6eaa22c03cb18e038","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.095","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3d3f4257a2807d7f460530c92673e11506ce9cb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1353,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=15;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1353,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7da641d323a8b845966c","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6d90dc1cc0401bceb0125ee","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.063","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd64a4df1589097a1dece4b339b4eb275234ed55","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1356,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=18;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1356,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.063,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-92025567d262d111a976","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e5521719c8b529123e78d2c1","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b13bcb9abbcb8a3eb83c48fac3c951a52a95f3c5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1344,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=6;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1344,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9daeb993cefd03d2369a","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-89d8a40bb5c0095db15ab2b4","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.136","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b13bcb9abbcb8a3eb83c48fac3c951a52a95f3c5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1347,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=9;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1347,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.136,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a427db77d11b82226ae","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e97176c852349a176f044238","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.135","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e59784b8e9d6e347874654899b64b81b4f1fd40","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1348,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=10;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1348,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.135,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-906a2493f450da1ea516","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1c46707e4d1c97cc6eba0741","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.074","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d4beecc719782af28c54689a3807cea1e1f7e48","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1354,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=16;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1354,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-39602cad61e0c0a46fe5","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4cb67d2bf8f3cc0b58f10cb7","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.169","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b805f68a5842bbf450d1438420ad9d97c6f2835","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1345,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=7;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1345,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.169,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-13df387838e0ab67f0bd","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-40af1f564b92030fc5417c9a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.173","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c47605036ead245c008913bc5824de46c6f1427","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1343,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=5;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1343,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.173,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-72761697e68ea2916d2a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9b1df0fcc3cb0f4fbfe7e2d5","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.193","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8748315c39d1c440e50edf7bdbfc9bc143441dc6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1342,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=4;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1342,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.193,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5f28f23668fb1d0166b1","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1a334fa316da81d1837018f3","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.138","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8748315c39d1c440e50edf7bdbfc9bc143441dc6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1346,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=8;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1346,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.138,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-93838158d9a5bd0d3041","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eb58b6ed4d3228237378207f","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.217","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8748315c39d1c440e50edf7bdbfc9bc143441dc6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1341,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=3;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1341,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.217,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6b8c0dac15d70337e932","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4f3dc4ef2c63847cbd35a3e9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"935cf3e3ad5224d6f5b0b04d5e78dd654411208c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1340,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=2;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1340,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2304f313ef51fda0e3de","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-13933aa7ce98a97bb405328a","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.132","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2925f0944f095f7fb8cb08daeb68d99ecdcd9bad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1349,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=11;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1349,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.132,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-fb79989e52543bb98c37","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-af824ce917d90ba80917985f","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_unverified"],"rank":null,"rawValue":"0.069","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"64c808df1b2d982da169e381315323340a6ae20a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1355,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=17;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1355,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.069,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4d5c6c2dfb8285c78fc5","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe34fef512bdb793b47a20dc","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.12400000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"013409308c458990176d82960959c65ec9467ff1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1350,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=12;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1350,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.12400000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cl_bench_life_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6451b7b1e85b31c20a9c","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be9b21ad1adcf8031248dcf6","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall","modelName":"GLM-4.7","modelRef":"zai-org/glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.109","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f828d2c001c0c67248a27165e9d2eee11035a928","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1352,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cl_bench_life_external.csv:row=14;column=Overall","sourceMetricId":"Overall","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1352,"metadata":{"epoch_file":"cl_bench_life_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.109,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-09-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1ff6e1e03163c3548c04","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1159f02ccee2cb983abadb75","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V3.1 Terminus","modelRef":"DeepSeek-V3.1-Terminus","observedAt":"2025-09-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0171428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6521c40bf23e361d21dc6a590c2c2b95a7917a2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1438,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=77;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1438,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.71428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-83ad0a5eb562ad51613b","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4aba9b883c3e45e9e3a0f740","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0285714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"43b8e739ca760638c19396b6dd8b875ef45fad5d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1433,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=72;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1433,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.85714285714286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b694bb717991e31287a9","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8be536e4a29886049a0a4d32","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.00571428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3a5b2ff71c472c8f0983df0e732860c4767c31d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1454,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=93;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1454,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.571428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c35018e2e0d0b39a9bf2","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d9148f15daf214e3a3c72a0a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0371428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03da8c1247f69488dd9a858c51cc04257517f014","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1422,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=61;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1422,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.7142857142857104,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ef6aa99a0da69b73ed14","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-063d5043bac8039a73068c18","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.2857142857","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d918d2d7de6e8de90efaea52a0bb61faab13598","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1369,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=8;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1369,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.571428570000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-877c3e2ea78df8fc45f7","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2874b36bed485426b007fd7","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b60ac9e1c14e0cb8dde0b4a6ea660f707b268e9f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1470,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=109;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1470,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ed33f5521ac56726791","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd5a88f894b921bfeb495e68","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b72f3189c0e3d62db41e3531abe36f3fbb117231","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1455,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=94;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1455,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b0e57f4db7613952ccc","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-47df187afed1f4b5c5937b20","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"937eea81a714114f58fe9349ba5129c2eadb3ba5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1400,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=39;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1400,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2a269b4bd8baa4148a3c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f6c5c2e16ddd244d9c7b541e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.20857142857142902","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef67c1996c99c7f0e3756faa3df3b302225d09fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1383,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=22;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1383,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b87e04af68e930e95992","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0054761eaeb01936f3a26fc5","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.282857142857143","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5329e3ada75fd056198cb801f188f9768562c6","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1370,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=9;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1370,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.2857142857143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2a1c1ee4fc8771eaf13","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2bf5c7588fa74039e56045e1","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.231428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5329e3ada75fd056198cb801f188f9768562c6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1380,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=19;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1380,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.1428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1f8c31810ec06f3096b6","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d807f78b105712d00b6583bb","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.291428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5329e3ada75fd056198cb801f188f9768562c6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1367,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=6;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1367,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":29.142857100000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f6e587ad4035bf76d8df","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0e84bc6035f8d348df1ff5aa","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_medium","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.268571428571429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5329e3ada75fd056198cb801f188f9768562c6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1374,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=13;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1374,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6b62422da69965347912","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-340dcfa87c2c0c548cb4c43a","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_xhigh","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.277142857142857","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d5329e3ada75fd056198cb801f188f9768562c6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1371,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=10;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1371,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.7142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2a1bca95e0d8160e919d","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-35a8f8ddc2504d8f8091ed66","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0114285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ad2a886c1f7b8acbc09a24f8ef415691ef1bc60","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1446,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=85;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1446,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.1428571428571401,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-361a9ed3bcee94517924","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ecec90d521a66702eefc8bf","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0314285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81d8272be7d7c0c9ed7f50bd1d3f9f4902015588","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1427,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=66;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1427,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-54f5dc8b8abb928236f4","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fb97fb4cd24f80b10aaf695d","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0114285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fca607d0d9fa288005c89968ce6cbd6cc124d3c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1448,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=87;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1448,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.1428571428571401,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b055b3906f938dc1daaa","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d63d36229d4f8024861042c","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.168571428571429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fca607d0d9fa288005c89968ce6cbd6cc124d3c3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1390,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=29;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1390,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a63277c8a15354b7a33","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f4746213142916fc895a3c43","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.165714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dfcf4a08d9ad5ddb9d8a9712d01f7e13cc7a5820","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1392,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=31;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1392,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.5714285714286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c2c342fe82603d86c5b9","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-401f0284df6df2ba0232578b","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0342857142857143","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dfcf4a08d9ad5ddb9d8a9712d01f7e13cc7a5820","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1425,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=64;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1425,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.42857142857143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5d0a76c8e0f91c0e83cc","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c78d720c462c092d80cdb3fc","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0714285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dfcf4a08d9ad5ddb9d8a9712d01f7e13cc7a5820","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1414,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=53;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1414,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7eeed2207d28cfabb35c","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c02d64a6a7af840f9c898428","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"345a99de32bbff42d64de35d547b49c549abfb18","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1405,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=44;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1405,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ebd9d27a9fc8a51eea7a","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d0952d3c1319e9ba862ab5c0","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.128571428571429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"345a99de32bbff42d64de35d547b49c549abfb18","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1398,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=37;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1398,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2029345b02e3683a64bc","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4823853f00e4ef94403dece","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.011000000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30ad8c351dfd7bba119532b3ddb0c8624b392a35","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1450,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=89;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1450,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-85f0bb9e745ce52415a7","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e30867950b38a6c52a16a374","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fbc7b7e9a76c4881429c30999b1ca0a6b844617c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1436,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=75;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1436,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5202036d243286a7edef","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-590c550c9beb7b7f25e3ecb2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.177142857142857","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"faded99987d9858bc6dc05f9c1eb72bf58a1f865","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1388,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=27;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1388,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.714285714285698,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-15b319bc6caff80055be","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20c0c0ba7d41105c95b94495","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf18b5f83e259c97deb3d7b02d5860d38228d60f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1482,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=121;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1482,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-901c7ee31d5a3d48c731","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1874a0da2dcd438e40e6b712","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.131428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d3219fab16c70d16254b3fa503a4afee368d4662","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1397,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=36;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1397,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.1428571428571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d1662b6f4c40083e2bf7","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0981a795aefd43cb32250a51","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.105714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"233ba22604f2c0d52ea49ba1aed008ac1e47903e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1402,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=41;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1402,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.5714285714286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-76032a5b901fa2dffa57","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-344259f8d227755512ef1b25","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0171020408163265","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49018a0222ce14c4be1fd5b2630115ffd23dedce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1439,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=78;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1439,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.7102040816326498,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec41b38edd77ca0d5172","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1ec86b0af8cb9cefd8643761","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0457142857142857","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"98e2201e0b005c074a71d3245390ef134359a32c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1420,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=59;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1420,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.57142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5f4afc5b98dda88785eb","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b5739bd28f013c6e98d3d396","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.20857142857142902","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9366fa644b2de01ee3e1d713432982f1374f5e11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1384,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=23;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1384,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59354a07eab688fb55b6","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7908b2fb551fc5b6f86e05b2","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0314285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9366fa644b2de01ee3e1d713432982f1374f5e11","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1431,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=70;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1431,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6be3287dd5b6a5e08c26","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eb07964801bbc4a482e7358d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.126","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffec66da43b88aca430ff2b43313c1bfa79c941d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1399,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=38;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1399,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3f0cf564d358069753f2","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fdc5a8e80a380222e8fbbbf7","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffec66da43b88aca430ff2b43313c1bfa79c941d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1466,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=105;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1466,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cb2f19e766e3e09ccd18","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98ed20419d96e716ae818177","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c5b1540933e88081468ac35fac37835c56f6d54","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1496,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=135;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1496,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b48e74645a2b7be21e1","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9704985a8223fa9c5fa728c5","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0485714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"caed8c2ca36495a28480e441c394caea5ac36b7a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1418,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=57;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1418,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.857142857142859,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0a5f48865d44fa4771f5","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32a4817b08f9ce71ca1713fd","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23428571430000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9281a38f99dd789a2e0cb5245cf8d53d9e4b06cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1378,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=17;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1378,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.42857143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-011883e14674710f33ab","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-95d4b802736dd376313e7b35","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"537ea46dc7d76b1da7c04b7f1e0150b0e32fa7a2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1403,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=42;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1403,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4090ccfe91799e591050","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-298f0598adde999db98a9559","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0925306122448979","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a217de0fe7232cf61a1e3727a4f1a73a282cc25d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1407,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=46;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1407,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.253061224489791,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b08138fc9ce21e574edc","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa601a4193fd60fbc5e0d452","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1de7a5e635c60aaa81715a13adbcbaec97a7acde","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1365,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=4;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1365,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-05-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-94651232c11aaf0bb438","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b102a9172041ce4ead9251a1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d98162fa32a19e02aba7eccf52ff5fe58f2f4071","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1499,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=138;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1499,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8df8f1d23711cd1260c3","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f1fe7d3b56a8ef1a5a0a14e4","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.305714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9bc4dacdc5051f4d379d151f2ca48e9bf5d6780e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1364,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=3;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1364,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.571428571428598,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-07c39535621225e63166","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-224ba0d9f79b652994d162ba","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.254285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d98162fa32a19e02aba7eccf52ff5fe58f2f4071","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1377,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=16;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1377,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.428571428571402,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc4becbd2cccb380a330","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1cede2452e77b1c2d636cf83","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d98162fa32a19e02aba7eccf52ff5fe58f2f4071","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1411,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=50;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1411,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bb3e58687488c88e8724","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-22b28089576931d76aca46a1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.18571428571428603","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d98162fa32a19e02aba7eccf52ff5fe58f2f4071","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1387,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=26;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1387,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.5714285714286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-668e9a8cdc6eb5e97204","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0784d2bb74ff456da260d1fd","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0142857142857143","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d98162fa32a19e02aba7eccf52ff5fe58f2f4071","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1441,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=80;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1441,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.42857142857143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad9d3555727a54343606","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7f39fccf6b77fd8d12401e6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.271428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d98162fa32a19e02aba7eccf52ff5fe58f2f4071","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1372,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=11;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1372,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.142857142857103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3f300055ffc30f8445d2","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7249793bfa12922b577941c6","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.165714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a304298fca361f50fa3ac69fa45cb607914195d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1391,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=30;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1391,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.5714285714286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d08d0743e8d33c94a24","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9818494cdb9b8e3345311227","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0257142857142857","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a304298fca361f50fa3ac69fa45cb607914195d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1434,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=73;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1434,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.57142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-817ebbc987613c02cef0","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1c9fd584727cddc1b6c0128","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.206","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a304298fca361f50fa3ac69fa45cb607914195d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1385,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=24;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1385,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.599999999999998,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-881fb3484a93a1a57d4f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-21ca7816cb14ea941843c49c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0485714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a304298fca361f50fa3ac69fa45cb607914195d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1419,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=58;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1419,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.857142857142859,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6a3b4e7dbe098a3496fb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b453eb2d39d9345b4f182f4","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.00285714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a304298fca361f50fa3ac69fa45cb607914195d","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1464,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=103;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1464,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.28571428571428603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c0d43823728a43c53631","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5fa9bb581e07af45639878d5","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.205714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a304298fca361f50fa3ac69fa45cb607914195d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1386,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=25;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1386,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.571428571428598,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d08ac37be7835119a32","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c316650f84c3b544c9a3b8f8","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.257142857142857","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c49848aa1e5f12525697fa3331fbec0a0c9a429","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1376,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=15;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1376,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.7142857142857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e7ba863d8c186a3b3dd9","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-627ec9cbfe866c34ff1f7215","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.148571428571429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c49848aa1e5f12525697fa3331fbec0a0c9a429","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1395,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=34;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1395,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.857142857142899,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-924f48c0cc476f2ab060","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ccc970092bcde6cf9d7fa1f9","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.323","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c49848aa1e5f12525697fa3331fbec0a0c9a429","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1363,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=2;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1363,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.300000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5d8c4cfe6285eb5f34a7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2625fce1c45e65ef5a762d6","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.228571428571429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c49848aa1e5f12525697fa3331fbec0a0c9a429","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1382,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=21;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1382,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f4c3bda5203ffd7100e8","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fae19dbb41a98d31057b40b3","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0514285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c49848aa1e5f12525697fa3331fbec0a0c9a429","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1417,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=56;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1417,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-34f3804dfd1c41b24c38","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-50acae49aa956931819cfa05","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c49848aa1e5f12525697fa3331fbec0a0c9a429","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1368,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=7;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1368,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.599999999999998,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dda1373a6b49fbac2bed","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba982c12d7ac5fcebe8ad13f","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.228571428571429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf930cb5483f5b782fc36463a28c5707f1fbb0b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1381,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=20;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1381,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.8571428571429,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-391bfaf6beb75464a3fe","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dc4c95816752098656f7f36d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0942857142857143","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf930cb5483f5b782fc36463a28c5707f1fbb0b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1406,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=45;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1406,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.42857142857143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce97e7f8694159b2f2e3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-337fe229bd3904e2636c7273","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf930cb5483f5b782fc36463a28c5707f1fbb0b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1366,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=5;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1366,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5000d3c7c483125c242d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0609fc2481cfd284852c34e0","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.174285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf930cb5483f5b782fc36463a28c5707f1fbb0b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1389,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=28;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1389,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.4285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-37e65464396772f785e8","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9cf01d6ce3b09456a5405257","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf930cb5483f5b782fc36463a28c5707f1fbb0b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1437,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=76;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1437,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2b62fdd2f6d39a7df366","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2e880381f3e286c3410618f5","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.271428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf930cb5483f5b782fc36463a28c5707f1fbb0b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1373,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=12;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1373,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.142857142857103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-48d93593126229111a9f","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c19b0aaf3ec9e6c3edf4459f","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.154285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7a5f8aa014bd7510d440eece9781e6d913a79df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1393,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=32;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1393,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.4285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d07c06aa33cdec671a25","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b060eab6ae1b9cb74ea11da4","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0314285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2601413fc49473c6658392c490d28225b94e51d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1429,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=68;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1429,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d0074fd985be8212fe6a","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a5d1cf6b8f1acc16d09c0664","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb75d623a136af3fd78757ded3c4f14243b18cd6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1412,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=51;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1412,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4dcea818c73cb44f198a","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86f60ee6f3798adedc268930","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"123cf6a4b5d87e5507851c08cdb87435880e15d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1404,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=43;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1404,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-532362054b0a3d7604c2","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3fecaeae3e7d329b6ab7fb65","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.0314285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9bef2d7a8e7b56b8c8a27ffdc3d3f8c44c142076","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1430,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=69;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1430,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee1259763ba451f89863","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d5c82478bcb864507003f958","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.234","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9bef2d7a8e7b56b8c8a27ffdc3d3f8c44c142076","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1379,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=18;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1379,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.400000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-6f5f5552f9e2a31f9d6d","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4aa98df35deced777d7849a3","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd1d0c1c518b94e9b19655e8fadf07bcc07fde86","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1472,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=111;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1472,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-130cb1f102eb54e4bb6a","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77a3d7905fcb886b7a616d8f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent"],"rank":null,"rawValue":"0.0371428571428571","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46fa170bc7c608ac2c7b0f17073ecc3075715fa3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1424,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=63;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1424,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.7142857142857104,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-19d3d403a7f7c20e0463","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c4561c09ef66404337c01ce","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent"],"rank":null,"rawValue":"0.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"092c35affe1995bc7e55fb0892f7b6b69fe678f7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1421,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=60;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1421,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-02-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2de45c984f9e58077817","canonicalModelId":"qwen/qwen3-coder-next@2026-02-03","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e03fe337cf86b3bebcc36387","mappingCandidates":["qwen/qwen3-coder-next@2026-02-03"],"mappingEvidence":{"confidence":"high","note":"Epoch exact Qwen3-Coder-Next spelling.","registry":"model_aliases.json"},"mappingNote":"Epoch exact Qwen3-Coder-Next spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3-Coder-Next","modelRef":"qwen3-coder-next","observedAt":"2026-02-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f538d9b6237e190bbb97166d0d87aaf70624a37","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1491,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=130;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1491,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"","release_date":"2026-02-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e7f97679c3ac16d2540e","canonicalModelId":"qwen/qwen3.5-9b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-10c7a6eaa5632b13ec94966a","mappingCandidates":["qwen/qwen3.5-9b@current"],"mappingEvidence":{"confidence":"high","note":"Epoch exact capitalization for Qwen3.5 9B.","registry":"model_aliases.json"},"mappingNote":"Epoch exact capitalization for Qwen3.5 9B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.5 9B","modelRef":"qwen3.5-9B","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.00285714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"184f79b3422715da65e5eaeae03376993111bc9c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1456,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=95;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1456,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.28571428571428603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0ab42afef9765ee5e7fe","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-55540e32dfd6ea8ea272d071","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.00285714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5ab405fcdb340efbb1a2227b190e8c1da8e2737","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1457,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=96;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1457,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.28571428571428603,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d3b91b861720f80db232","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-99a8ead606b2a73ac59d546a","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0285714285714286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fded4f9a8d75f9b8293445a346f72605568c2638","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1432,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=71;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1432,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.85714285714286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-78f341909e310a0bd4b4","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9d41c4e5a64e8947e021dd82","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.134285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21af0a3efe47064e0b5f604665cbd249f2314122","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1396,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=35;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1396,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.4285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-critpt_external","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-57647ec6904d5932fd8c","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ad89304a1574736ab354c403","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0914285714285714","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57164e6b4471c96901c009f51793018d8f234f3d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1408,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"critpt_external.csv:row=47;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1408,"metadata":{"epoch_file":"critpt_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.14285714285714,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1c108db6ebc3d3a8d64f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ce3da02e40d2dee8ea5d936","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f346f38044bd611b026585d8384c21fce0f4f9d5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1504,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1504,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6c9692ce60d366cdcbce","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20724793eeccb3a367f69b1a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.621","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f346f38044bd611b026585d8384c21fce0f4f9d5","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1542,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=42;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1542,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.621,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-568aab8bc898403dbf4d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b30ff2298312767a565c0fb","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.705","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f346f38044bd611b026585d8384c21fce0f4f9d5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1502,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1502,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.705,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-21107d52df9730cba500","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6071a317f09a227daddd8518","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_medium","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.652","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f346f38044bd611b026585d8384c21fce0f4f9d5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1536,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=36;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1536,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.652,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c22c86adcf54e2e258c5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ee08db86c10c916b3f989cc7","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_xhigh","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.684","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f346f38044bd611b026585d8384c21fce0f4f9d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1503,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1503,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.684,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d15fe05cbb3a2eba0035","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df8f4478b9eb1c07fd30939d","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.594","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28c85029f1666fd4072dd38b1578ac29a3c9d24b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1511,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1511,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.594,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-913a6be2588f5b13ef4c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b21469e21158801c245f7a53","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_low","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.483","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28c85029f1666fd4072dd38b1578ac29a3c9d24b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1523,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=23;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1523,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.483,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-95e7402797d410014971","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cbe17d566ad9d0e2e6247e2e","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.648","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28c85029f1666fd4072dd38b1578ac29a3c9d24b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1505,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1505,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c1f7920c5f5e9f5faf38","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9df7713fc5b988b3f5f83179","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_medium","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.527","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28c85029f1666fd4072dd38b1578ac29a3c9d24b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1517,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=17;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1517,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-12205e8ddf891554fb4d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3819cda4be484343d54b850","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.616","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28c85029f1666fd4072dd38b1578ac29a3c9d24b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1510,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1510,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.616,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bad68eb97dd7909a69fd","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-54158735bed7bab4fcb35fb6","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac6e7106c4470c65df627a546b2fcd2daf2147b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1513,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=13;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1513,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.58,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32682c54cdbbe744c680","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d24b4c21c8cc6d33ae4e62fb","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.531","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac6e7106c4470c65df627a546b2fcd2daf2147b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1516,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=16;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1516,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.531,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-12d6cad21091bc45a6ff","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c4c10e0ecf7604e6d16951e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.623","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac6e7106c4470c65df627a546b2fcd2daf2147b7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1507,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1507,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b4823be0a477b5d9d3a5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33303bc7beca25c898c9206d","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.561","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac6e7106c4470c65df627a546b2fcd2daf2147b7","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1514,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=14;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1514,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.561,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e8e79a1790e5788adc00","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-36e4ee07a90407eff2a552d4","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.594","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac6e7106c4470c65df627a546b2fcd2daf2147b7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1509,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1509,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.594,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-794d36dc01632436ffae","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-607cb45d909e9129736624ff","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49918f75eb48e73a99df1e4a90f996b39380db7f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1532,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=32;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1532,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-02af7076555977bd983e","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-814deee0156b84edfcdcb256","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.628","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49918f75eb48e73a99df1e4a90f996b39380db7f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1541,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=41;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1541,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.628,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-471b2a1a99d690644a9f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-302fc9375aafe295dc104aa1","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7000000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49918f75eb48e73a99df1e4a90f996b39380db7f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1530,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=30;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1530,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-017528c11dc108f0ac61","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1c969cb7899497bc040dc3b","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_medium","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.643","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49918f75eb48e73a99df1e4a90f996b39380db7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1539,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=39;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1539,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.643,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-883e29aa19ab22f7fa4c","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f51c3636a0ab9d98e6db757","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_xhigh","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.693","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49918f75eb48e73a99df1e4a90f996b39380db7f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1531,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=31;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1531,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.693,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0933820b37845460e77e","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac1c6b71c3f1f92362f85ed6","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.488","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f2442e1aa61c840aaa0145ec9c33428ac3bc24b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1522,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=22;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1522,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.488,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1e2f4798d2f315994b10","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-330c0a7959e7b1588ab1835c","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f2442e1aa61c840aaa0145ec9c33428ac3bc24b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1520,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=20;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1520,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eaa80c30f7ade962f41c","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-505e0caa9554ca8a3a5df997","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.46","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f2442e1aa61c840aaa0145ec9c33428ac3bc24b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1525,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=25;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1525,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.46,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aca0a2c15cab31c0c35a","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fb294cf8b6af1eb111aa9757","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.569","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58c88b400b7efd6fc1711e54dd8e894fede9f741","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1552,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=52;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1552,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5b67fa714dca506ceadc","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-38f5af9b2af8291284278874","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_low","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47700000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58c88b400b7efd6fc1711e54dd8e894fede9f741","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1565,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=65;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1565,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.47700000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d6584f0e15fa37bd40e9","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db367adfb7dd736266669201","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.615","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58c88b400b7efd6fc1711e54dd8e894fede9f741","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1527,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=27;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1527,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6502ceec902ecd7bb971","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bc6292596c4648cab0b8e5a8","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_medium","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.524","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58c88b400b7efd6fc1711e54dd8e894fede9f741","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1559,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=59;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1559,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.524,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fd996af6c5544dce0a0d","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b0eea32127933f627df3c7ba","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5870000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58c88b400b7efd6fc1711e54dd8e894fede9f741","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1528,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=28;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1528,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5870000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6267e9cf8674b5f73356","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba38abe1e4ac2b29e97873c1","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.498","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35083d667fc515e8e6fd5104d6e7cdb603c3c37b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1519,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=19;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1519,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.498,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce05d9075359093d1abd","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e7506baa5e04bea25f7274ff","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99bf955e3986cade245ab0162ec57ba353fa748b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1557,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=57;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1557,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e91a241ca6e97a957ee6","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df19af12ebd2e398cfb78385","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.474","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99bf955e3986cade245ab0162ec57ba353fa748b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1566,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=66;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1566,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.474,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-592d246a6a382e200156","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0939581e4df4b286af90e7d8","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.512","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99bf955e3986cade245ab0162ec57ba353fa748b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1560,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=60;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1560,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.512,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4752c8d0d82cc51a9cc3","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7ef937f8a04a893435235675","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.616","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da06a1faad4ac1cec5dc641fa3f26be0caaf85df","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1543,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=43;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1543,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.616,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4b0619c23b0488b43543","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d19e4b9e199bac6259fc924a","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.538","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da06a1faad4ac1cec5dc641fa3f26be0caaf85df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1556,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=56;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1556,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-52e8ee14dcbdba2851ab","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e4fc7c747a6601354f019fc","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_medium","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da06a1faad4ac1cec5dc641fa3f26be0caaf85df","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1550,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=50;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1550,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b47b68963405cb958a99","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03697f8c8056c8675ea536ed","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_high","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.515","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cecd31bf0857240022397611b6a0c0971c749c8b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1518,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=18;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1518,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8aa6dcc584fde35fef85","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-852df93f5f7ec6b3394577bf","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cecd31bf0857240022397611b6a0c0971c749c8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1515,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=15;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1515,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ff97141476ca53512a73","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-04faf7dde8d20515df63cc29","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.584","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c3916ec341c5ade401c7dfe94b171d1750a6239","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1508,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1508,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.584,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-330bd02a4eb0e44c07ea","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9873a1eff70d434ee8062c83","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.466","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c3916ec341c5ade401c7dfe94b171d1750a6239","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1521,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=21;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1521,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.466,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ecdeb43c7d4dc4f8fdbe","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ccebc171b86b1813265c447","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.538","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c3916ec341c5ade401c7dfe94b171d1750a6239","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1512,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1512,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-624a59ca6b7de936e292","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-38d00b5a1dbe3d0b38985f49","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.584","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c3916ec341c5ade401c7dfe94b171d1750a6239","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1506,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1506,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.584,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-02e0ac0059c8c3515c0d","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f622390ad3c45f722282ad9c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.568","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47207f8d7b6303f497f7fe2d8828ee419955ad86","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1553,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=53;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1553,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.568,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0bce089b528626c1d627","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5705feeac2ee7f037e24817b","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.376","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47207f8d7b6303f497f7fe2d8828ee419955ad86","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1568,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=68;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1568,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.376,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-be5b3d2d16f972f2f692","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03cffd860ba19482f4690675","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.611","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47207f8d7b6303f497f7fe2d8828ee419955ad86","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1544,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=44;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1544,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.611,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3048c65fcfdeb45a93a3","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0114d68764f1a6afcfe8cf11","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47700000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47207f8d7b6303f497f7fe2d8828ee419955ad86","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1564,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=64;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1564,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.47700000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8c84a116642e916755cf","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-faec571b28a64e5dc4693775","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5770000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47207f8d7b6303f497f7fe2d8828ee419955ad86","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1551,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=51;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1551,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5770000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d202881b1b3c9d762470","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d0951e8a6a2dfea1bd2e4637","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.635","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c2d458cd6dc4e6c730d39f8befd31b392a2c185f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1540,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=40;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1540,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.635,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-48baa135c8089254289c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8be51d18d70a927bdd7187d6","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.526","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c2d458cd6dc4e6c730d39f8befd31b392a2c185f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1558,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=58;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1558,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3756d5f2b672647c10c5","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0e11c2a2fcb9462a89fac733","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.672","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c2d458cd6dc4e6c730d39f8befd31b392a2c185f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1529,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=29;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1529,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.672,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2c7214d2a9a3196d9ce4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c9b6965b0dbbbc34851c22c0","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c2d458cd6dc4e6c730d39f8befd31b392a2c185f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1547,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=47;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1547,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f98b331acd99d1851641","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6cf71149871a85664ddcb8d1","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.645","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c2d458cd6dc4e6c730d39f8befd31b392a2c185f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1538,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=38;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1538,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.645,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-928b3369dd4205e58fbf","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f02044307f619f42d98af64","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.542","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"203898d2b740cc236cd54e0758c91664aeadd221","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1555,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=55;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1555,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.542,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c431278513457325ec47","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-411579c278d01373f3b663e9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.469","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"203898d2b740cc236cd54e0758c91664aeadd221","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1567,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=67;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1567,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.469,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cdb462f56b4d37b74538","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3a68530577e4f285643f889","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.649","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"203898d2b740cc236cd54e0758c91664aeadd221","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1537,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=37;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1537,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.649,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f5b7022e23436f459368","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c58cd2564392b72ceb1f38aa","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.503","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"203898d2b740cc236cd54e0758c91664aeadd221","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1562,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=62;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1562,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.503,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-10af71d58ef293580942","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba678e10845b38d3125fdb9d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5920000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"203898d2b740cc236cd54e0758c91664aeadd221","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1549,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=49;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1549,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5920000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0a28afe0cb9a07df3a64","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b181d2e4ad41a54e0e1a959b","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6990000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c514b0eb8eb99c791bc84185d7b7711c142ab292","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1534,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=34;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1534,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6990000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1a437e4e44bcf0a32dae","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8ed5fe0ef2c6406d68891e10","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_low","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.61","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c514b0eb8eb99c791bc84185d7b7711c142ab292","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1545,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=45;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1545,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.61,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3391c808cc6927279bca","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c168193b74508e556625aa52","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_medium","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6709999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c514b0eb8eb99c791bc84185d7b7711c142ab292","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1535,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=35;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1535,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6709999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2b5f6e4e932c8b4ce1e5","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b6e28ea15621975307103f18","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.708","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c514b0eb8eb99c791bc84185d7b7711c142ab292","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1533,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=33;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1533,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.708,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4f7cb9787ad35b2725d1","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-acb0cbfac40acf4e8fa5502c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.319","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85b66878c3ad0cbcf6fcb77031d0aca1f85a6cc3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1526,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=26;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1526,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.319,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1c6dd934b99443abab6a","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0e649c38aa698a45d60ab2a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47600000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4e19552d00ae9fe742cb0c8ef6dc8ef2da7afc91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1524,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=24;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1524,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.47600000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e09d2be29ed7bf8ab17","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dc1ce9b1d9bd5f2275402742","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.49700000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1e11417f2337b08d34c13f52bacf1d6520ebea3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1563,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=63;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1563,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.49700000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77eaded1cb34578f1a2b","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9be92eea5cfc86f902571533","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5970000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4c883541fd1c335b43e5fedacab6955234de17a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1548,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=48;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1548,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5970000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a46505b7bb4bb937cf1c","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b19f9f84708e9869bdd63155","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.505","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4c883541fd1c335b43e5fedacab6955234de17a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1561,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=61;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1561,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.505,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-cursorbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ac18a1bd401d7a074a30","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-44a00db99d686293ecca6a7b","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.608","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4c883541fd1c335b43e5fedacab6955234de17a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1546,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"cursorbench_external.csv:row=46;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1546,"metadata":{"epoch_file":"cursorbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.608,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1516dc5eedc3b84f506b","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3e71e8bcfaecd53f4e8a41d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.351","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e833aa24f5466279d72aefe647594c28600185fd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1609,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=42;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1609,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.099999999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-af31e6859c775e172e4d","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd763ee278375f6f95a64544","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_low","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.455","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a4a5e417c44e14481cef47523ffbb7630e7e94f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1597,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=30;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1597,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-325c0fdfbe5d9ad9aae5","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-458a24fae7d0bb8f67049e57","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_2K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.46799999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f4265857037a67ac132c4abdd6f72b311f7bd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1594,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=27;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1594,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":"2048 thinking token budget","organization":"Anthropic","release_date":"2025-05-22","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-879409a9bce707bebe1c","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ebae4740211513c83abfc630","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_high","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5479999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b668a2aaad8730f5216f4ba8178883f19748485","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1571,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=4;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1571,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.79999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-65b7ca4c6b8a7fe5db7a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b6463caf36b5116a2d3a6c34","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_low","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.537","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b668a2aaad8730f5216f4ba8178883f19748485","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1573,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=6;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1573,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7badd0ae8f4a95745d88","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e9ff875369e0f7b362ed5d4f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5529999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a541727e65cb6c27a50e154673d12483625fb048","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1569,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=2;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1569,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee795b40e6b96f453f6f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08a3be5da7b9620ea600c0a5","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_low","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.514","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a541727e65cb6c27a50e154673d12483625fb048","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1576,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=9;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1576,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":51.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d2507100157f4af3d4b6","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c75950b4f5b2e13c0cf44622","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_medium","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.532","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a541727e65cb6c27a50e154673d12483625fb048","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1574,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=7;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1574,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c4f7b3dbbe13f4cf05d2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ff674b7731f1c913f6362284","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.502","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52f92e4a05ddfafcc2510590f80cc12f9d9323bf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1578,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=11;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1578,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a97d959dc3883af36f92","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b5ca72526bc78196c38fa7b","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.493","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52f92e4a05ddfafcc2510590f80cc12f9d9323bf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1582,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=15;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1582,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-953bd08aad2acddc5744","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-da431f1e00acf39ff8243c7c","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.474","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52f92e4a05ddfafcc2510590f80cc12f9d9323bf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1592,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=25;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1592,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-015c29843cd8c1aa3a47","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2d0b92b94c80b63450836b1b","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_2K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.526","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30a53131258be681bd0213ac9f28c3ddde6f9b12","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1575,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=8;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1575,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":52.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-253f633c6215a23ee753","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ba445b1dfd4b9074bab019b","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_low","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.47500000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30a53131258be681bd0213ac9f28c3ddde6f9b12","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1591,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=24;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1591,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-12e6d9f48ed528cf6ba9","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e7a22ad6cac30fc1532c2335","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.549","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cc29da9558cc1c41abcaa794d8ef07e00d8e43e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1570,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=3;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1570,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.900000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-39020393c74463a5784b","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b858e5bf0e2cbe3b54807b2e","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_low","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.504","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cc29da9558cc1c41abcaa794d8ef07e00d8e43e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1577,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=10;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1577,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b028ad0d8e1596498449","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e82138e6c38f8c9b90780f47","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.41500000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e98ecfc8b04466820990794b58a9168f70c63e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1603,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=36;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1603,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2a521cf556afcb117b48","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c50591a9fe6ad9edad0d1243","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.478","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7085f40375bcaaf7343c53b2e6801cd27140ff4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1590,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=23;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1590,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2d7cec56cbaae1af74b","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a417fa771310d78f1d0cfaad","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_low","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.445","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7085f40375bcaaf7343c53b2e6801cd27140ff4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1599,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=32;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1599,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-adb8160576a26e35e408","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-31a24cd611aa59bfb4259b90","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.48100000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27a3bb513179413699bd2db2dc374045afff8ce2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1588,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=21;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1588,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-678dc31d31a55037e354","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77e97b1827e62c680ba4403c","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.49600000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27a3bb513179413699bd2db2dc374045afff8ce2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1580,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=13;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1580,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.60000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e82702444d5f95000726","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80a9d616f60b5568d2115d60","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.48600000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27a3bb513179413699bd2db2dc374045afff8ce2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1586,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=19;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1586,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-baeebe10494af905bf39","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://drb.futuresearch.ai","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d529a252b7d36cd4cb231514","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.489","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27a3bb513179413699bd2db2dc374045afff8ce2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1584,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=17;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1584,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://drb.futuresearch.ai"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dfcbd51c0de54c587dd5","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-10bf061e3dd1fa82492e5185","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.428","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc491d24c3dc248f4ead159c3ba5b383bc17e375","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1602,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=35;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1602,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cc8fc4ce968d0b5fd954","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a07a7147843d1cdd4ee26108","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.41100000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7be265cb81be000537697202550087f47ff70837","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1604,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=37;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1604,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f58fda2abf87e65cb251","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7d7548a1e648458099da87e4","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35100000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3aeb84300765cfcf0f569806bf7c12d513f4065f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1608,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=41;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1608,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-020591aaa7658ecf5849","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b5b546d17a9000bab6bcec0f","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.363","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d561e1540fe2bdf18d1c43ecb12d2528e3d6bfdf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1607,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=40;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1607,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-925988ff7e7b9d61834a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8db43b102d4c9e4e6eda67f6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.54","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bc73f913dbeacd7132b1a7718816e238e4eacac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1572,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=5;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1572,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d9421cb33d844aa5e9e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6ac87e66399417da4c08bf0a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.48700000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bc73f913dbeacd7132b1a7718816e238e4eacac","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1585,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=18;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1585,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepresearchbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7ffc87d1ecef69c8283e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-51dc2973fad304c36fb4d155","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Average score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.49600000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bc73f913dbeacd7132b1a7718816e238e4eacac","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1581,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepresearchbench_external.csv:row=14;column=Average score","sourceMetricId":"Average score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1581,"metadata":{"epoch_file":"deepresearchbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.60000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7bf7c19b18c17f08f064","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-4006cfb8beeb423641c60d68","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6860465116279071","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa886c625fddac8f3a9a5128ee5d677b34ec88f2","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1616,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=8;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1616,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":68.6046511627907,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cb5e796d9c744399566f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-c422ee4a9ddc3f9abb7d1048","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5958429561200924","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa886c625fddac8f3a9a5128ee5d677b34ec88f2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1653,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=45;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1653,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":59.58429561200924,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dce0b16f91447b7f0c40","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-ac5c103ae7a241051afb9489","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6972477064220184","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa886c625fddac8f3a9a5128ee5d677b34ec88f2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1613,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=5;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1613,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":69.72477064220183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0b9e2252f60dab2b2e8","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-e1b549213bccbe346e48de55","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Fable 5","modelRef":"claude-fable-5_medium","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6536697247706422","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa886c625fddac8f3a9a5128ee5d677b34ec88f2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1652,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=44;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1652,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":65.36697247706422,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f1bc69c0b04d4eb46d91","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-49d3d2a30c32b4815c768c4a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Fable 5","modelRef":"claude-fable-5_xhigh","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6991150442477876","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa886c625fddac8f3a9a5128ee5d677b34ec88f2","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1612,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=4;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1612,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":69.91150442477876,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f97f99f4f6a7b72c64a2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-feefd05ef1a070c7a871977e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5176991150442478","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c099b88aa02300fc7bf62040ab2fdcdea580824f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1629,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=21;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1629,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.76991150442478,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-60f3cc728a501249058a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-c2b3f8e300ecdfd392838c04","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4079822616407982","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c099b88aa02300fc7bf62040ab2fdcdea580824f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1637,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=29;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1637,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":40.79822616407982,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1538b97f1aad2e828215","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-1051a00af89b1d3a0bda5530","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5897435897435898","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c099b88aa02300fc7bf62040ab2fdcdea580824f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1622,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=14;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1622,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":58.97435897435898,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-65079ff0c91bf37596d7","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-b0f4c707b0b30a21732b03dc","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.48672566371681414","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c099b88aa02300fc7bf62040ab2fdcdea580824f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1632,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=24;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1632,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.67256637168141,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad25ffdb82b8e715fa6e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-17887d18f4d15476f75b18b4","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5436241610738255","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c099b88aa02300fc7bf62040ab2fdcdea580824f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1624,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=16;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1624,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":54.36241610738255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b0e3fffa52040dffad46","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-66ff88e44a1fd79ca24c61e3","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7282850779510023","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32621d1eccbda8a8b8741e65b804aa8534454e8f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1649,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=41;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1649,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":72.82850779510022,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5e4f553e886f7099889a","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-fef2a4a08594905904be386d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5812917594654788","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32621d1eccbda8a8b8741e65b804aa8534454e8f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1654,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=46;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1654,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":58.12917594654788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-40ea211274190a48aab3","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-fe8c81754ce28181adcd6113","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7364864864864865","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32621d1eccbda8a8b8741e65b804aa8534454e8f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1647,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=39;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1647,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":73.64864864864865,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-966bce2275295cdb5a7d","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-01a745c8ae82987f1dcf33eb","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 5","modelRef":"claude-opus-5_medium","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6890380313199106","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32621d1eccbda8a8b8741e65b804aa8534454e8f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1650,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=42;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1650,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":68.90380313199105,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9522731142018dceb2e8","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-4b1f6639e2ecab355a42c83d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Opus 5","modelRef":"claude-opus-5_xhigh","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7315436241610739","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32621d1eccbda8a8b8741e65b804aa8534454e8f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1648,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=40;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1648,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":73.15436241610739,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a1944028b6dcb6be48b4","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-987c8c783b5684c335f69acf","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.29933481152993346","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b27515191690d83643dc3c3a6dfbb451240223a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1641,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=33;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1641,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":29.933481152993345,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-40caad943202b031ac8d","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-00eb0fd0b539700fb70fe3b7","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4823008849557522","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c8fd58a2868b38d9dae0107e242e1bafffc6884","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1633,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=25;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1633,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":48.23008849557522,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-17e52437950f7e090c52","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-6bf522dcf76410c351a8199f","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_low","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3051224944320713","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c8fd58a2868b38d9dae0107e242e1bafffc6884","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1659,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=51;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1659,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.51224944320713,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e5466a83d40efe603ed1","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-388b084c2b544ae00cef4e66","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5384615384615384","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c8fd58a2868b38d9dae0107e242e1bafffc6884","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1626,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=18;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1626,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.84615384615385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a68baa9c7944cf8d1ea","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-efa238ed92a5f4584024d50b","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_medium","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3977777777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c8fd58a2868b38d9dae0107e242e1bafffc6884","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1657,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=49;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1657,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":39.77777777777778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2894e9eae80ab3c12cc9","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-a8b1d2dae4ceee9fe4aedfc6","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.49667405764966743","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c8fd58a2868b38d9dae0107e242e1bafffc6884","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1631,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=23;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1631,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":49.667405764966745,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8a2e8b50a67cd1f85dba","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-c089a9a3df1062b31994a310","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.1172566371681416","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"12628500cd73edf0e42a02796e447093494e715b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1644,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=36;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1644,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.725663716814159,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8858f52d0b87a52c7749","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-846a602a2c2f1267f34630a7","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3606194690265487","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f319238dc3e8576a3f4625af409f7a4aa96865c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1665,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=57;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1665,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.06194690265487,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2fd034451fe24fdb34e5","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-c0850a5da54b97511a19f7f3","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_medium","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.37389380530973454","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f319238dc3e8576a3f4625af409f7a4aa96865c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1658,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=50;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1658,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":37.389380530973455,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d4ae8eed715dd31c8361","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-9268916a0092830c60e48793","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4668141592920354","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0af3993676fd4455b828e6a2e72309bd56c9f61","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1656,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=48;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1656,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":46.68141592920354,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2340b9be03e4245cb17","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-c54120a92c4d1ff0beb83f8c","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6526548672566371","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80b0a490da3c5b4a62a64e774b15b434cc9069dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1667,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=59;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1667,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":65.2654867256637,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-35a2fe79078b4d80a7de","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-8be0a9d2c82a4cecbe8e9b57","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5376106194690266","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80b0a490da3c5b4a62a64e774b15b434cc9069dc","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1668,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=60;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1668,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.76106194690266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4e0ef7c99b00ba62a7fd","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-984e672ad8c3d0bdac675773","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_medium","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6548672566371682","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80b0a490da3c5b4a62a64e774b15b434cc9069dc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1666,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=58;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1666,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":65.48672566371681,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-25559c9eef7ccc8b2fa0","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-34768d821cafec20ffed8ce7","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GLM-5.2","modelRef":"glm-5.2_high","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.36283185840707965","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22bcd16e217d8fca0acb40bcc84b555d18f1c6f3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1638,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=30;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1638,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":36.283185840707965,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-462cca6836b5902c1b9c","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-ef20c9f778d375b7e9d27640","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4377777777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22bcd16e217d8fca0acb40bcc84b555d18f1c6f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1636,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=28;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1636,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":43.777777777777786,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-393ae24743935c4bcd3d","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-b13ff96f4fd4df8970358612","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GLM-5.3","modelRef":"glm-5.3_max","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6895787139689579","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eea8e20eb8891be4e3bcaaf92e8e9ee88b244d84","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1669,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=61;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1669,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":68.95787139689578,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eb1df425f800509f823e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-1ff031f17e0ca00109b2b59b","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5176991150442478","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11827d75ba115be3e62b56485e53b44a3a5866c0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1630,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=22;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1630,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":51.76991150442478,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b1ce8cc713dd97e47a14","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-3e26486b908330a7b3e04510","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6438053097345133","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bb6a487c0a6ace2e3871a5c7d59eef73e95bddd","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1619,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=11;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1619,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":64.38053097345133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6bf0be7d295a1ab8473d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-dac87db8c75127ce84e6a384","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.26991150442477874","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bb6a487c0a6ace2e3871a5c7d59eef73e95bddd","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1642,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=34;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1642,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":26.991150442477874,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6eb41b9fa297bd27d4ef","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-e522eb3ea96be2633f714fbd","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5398230088495575","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bb6a487c0a6ace2e3871a5c7d59eef73e95bddd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1625,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=17;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1625,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.98230088495575,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-73a42c6d4d456c14e3a4","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-57ecee3160942a344e8fca07","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6703539823008849","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bb6a487c0a6ace2e3871a5c7d59eef73e95bddd","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1618,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=10;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1618,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.03539823008849,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59981b2d5453579a5092","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-28a2ae7bb1e7d4c0d767103c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4424778761061947","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1eb3b1f3e8e8f30105c55953c4072e3b84f6d6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1635,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=27;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1635,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":44.24778761061947,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04bdf9138fcd32b91f90","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-26470f3000a2bdf6bdcb248e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.015486725663716814","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1eb3b1f3e8e8f30105c55953c4072e3b84f6d6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1646,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=38;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1646,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":1.5486725663716814,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c2cf6f8a7c423493fb75","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-bb2638a6d5816563ca92d7eb","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.671875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1eb3b1f3e8e8f30105c55953c4072e3b84f6d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1617,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=9;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1617,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.1875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa47ea3946409b278e31","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-7d71cfd019ca25326cd215aa","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.11283185840707964","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1eb3b1f3e8e8f30105c55953c4072e3b84f6d6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1645,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=37;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1645,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":11.283185840707963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a15adeef632594f0209c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-1e9e65475e542a06dfa1dd3e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5685840707964602","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1eb3b1f3e8e8f30105c55953c4072e3b84f6d6","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1623,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=15;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1623,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":56.85840707964602,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-764b4f8b6736a42d0a9b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-5826ef81cdd691d50f75543a","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6940133037694013","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c05043aef27cda25e409de352396332dc1d7fe5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1615,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=7;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1615,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":69.40133037694012,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7337761c02023b0c9363","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-3ba37e445b3f805b0e298b7b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.45353982300884954","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c05043aef27cda25e409de352396332dc1d7fe5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1634,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=26;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1634,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":45.35398230088495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32a929f2fc8a694037a0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-241cf6362fb6190218f5a9ea","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7266666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c05043aef27cda25e409de352396332dc1d7fe5","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1610,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=2;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1610,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":72.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-970d3735c6a49d072126","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-270b3dc7d6300aba2c04117e","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6106194690265486","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c05043aef27cda25e409de352396332dc1d7fe5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1620,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=12;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1620,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":61.06194690265486,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-763dcc82adcafa133684","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-a69d821c31df6e781fad66cd","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7073170731707318","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c05043aef27cda25e409de352396332dc1d7fe5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1611,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=3;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1611,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":70.73170731707317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b48454248cc91e5b259f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-3301d03829d4efde8d520b37","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5376106194690266","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d787a9a263b8b4040622200186984ac29a08043","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1627,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=19;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1627,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.76106194690266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bc168ffb6be1d1261872","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-eac37568f1c8a9bbf330c79f","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.24053452115812918","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d787a9a263b8b4040622200186984ac29a08043","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1643,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=35;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1643,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":24.053452115812917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce7a10cdec20bde2f991","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-e32ee58097f7c552f2ed68d2","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6962305986696231","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d787a9a263b8b4040622200186984ac29a08043","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1614,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=6;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1614,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":69.62305986696231,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e9ea1315f00ba0b5389f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-b2cc7adea37d644577ccf7c0","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3511111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d787a9a263b8b4040622200186984ac29a08043","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1639,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=31;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1639,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":35.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77cf8a11c090b6590335","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-4a2880700ff84412f5ced15d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6017699115044248","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d787a9a263b8b4040622200186984ac29a08043","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1621,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=13;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1621,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":60.17699115044248,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a5e1110496ab60986f9","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-fa49913d440bf7532cdbc150","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5376106194690266","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"740834eb83e2883ba3670a84d404b05c834b6deb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1655,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=47;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1655,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":53.76106194690266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8cc2330d8e0f77fed7f7","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-673b43b063f3a2e7084d6185","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6518847006651884","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef63b551629ee5ccc07902e62e118a3b09517e7f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1662,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=54;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1662,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":65.18847006651885,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0cd598af4e57056a81e0","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-f1763c8bb34340c1e1d00532","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Grok 4.6","modelRef":"grok-4.6_low","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.41648106904231624","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef63b551629ee5ccc07902e62e118a3b09517e7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1664,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=56;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1664,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":41.648106904231625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f32712858c2767e324f4","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-bfe347ccce01c5792c70493f","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Grok 4.6","modelRef":"grok-4.6_medium","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6747787610619469","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef63b551629ee5ccc07902e62e118a3b09517e7f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1660,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=52;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1660,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":67.47787610619469,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6ea9fcc1d0f70b8cbdae","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-bdf3565f4da9e7193935c671","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6674057649667405","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef63b551629ee5ccc07902e62e118a3b09517e7f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1661,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=53;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1661,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":66.74057649667405,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-18e9b1de46bc923aedcd","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-956ce259f15ae719e29cb55d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3053097345132743","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68a6af6612e334dd5169b1f76013f64178305af5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1640,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=32;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1640,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":30.53097345132743,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fdb4da949f1ce81a7595","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-463e5f1dd24cdfd06d9f2250","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6851441241685144","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"264ad6bf4eff4fab9cab6af91a50bbac668a0713","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1651,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=43;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1651,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":68.51441241685144,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-deepswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-91e9d6b7bbe58900627f","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-4c0465286098c6cdd74fc847","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Pass@1","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5746102449888641","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a02598ab9f4199495bb66611f2df4e20ea58e9ec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1663,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"deepswe_external.csv:row=55;column=Pass@1","sourceMetricId":"Pass@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1663,"metadata":{"epoch_file":"deepswe_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"percent","value":57.46102449888642,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-08189045cc72a8f2f6c0","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-754e717dd113a0303340ee62","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.39280000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f3783495ec8a15e7e899a1fa59a71f66cbe94977","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1713,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=44;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1713,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.28,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fd37e8fee1fade5b3d3f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a241445c4a9f02a94783c3e9","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.055700000000000006","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"049d3606216abb5767a163f3ed4f73fd85787625","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1691,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=22;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1691,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.57,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d7229618e02c8568b21","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a39bba733ba5df1eaddbaf74","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1191","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"15f2ae1f4efc19ee455a336d49d7fc5a624efeab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1677,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=8;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1677,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":11.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7c7d075a41adfa5493f7","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4cc5fe33efa8f04d576e49a0","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0684","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a914876bbada3aa7d74ae463d29e55c395186124","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1686,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=17;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1686,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.84,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a984e950e3c06ff7f9b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7fa89aca60eb77dbfecf21da","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.076","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a914876bbada3aa7d74ae463d29e55c395186124","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1684,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=15;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1684,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8c6338817df663ecaa5d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6355637b9bceec706cf9ca45","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.23510000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8231aac5b3837a577160cd81c5fa176dac3a29ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1716,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=47;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1716,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.51,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d081e49d5bcd6eac4aa","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-34aff24e9a6786f13dc30e46","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.059966","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9213bfa96df857e1ea2cb4112ab406ef652d365c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1689,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=20;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1689,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.9966,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-57c79a28f065ac7120dd","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6d71036cd95d4c36e3dfd7e1","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3678","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25928fe4c2fcb6ebbb7eea0b6045ab2874b23fdb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1672,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=3;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1672,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.78,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e4e78ed4b2ef05480392","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9c622144d4ce69a7fd9ccf2b","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2541","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"532a91f42fd291629741afe31f195b72418065ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1715,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=46;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1715,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.41,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b79aa2c2b439767419c5","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f757a275822dad0ffe30496","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.10470000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da50dce0b797960ab7a58b494f1b542de24cd2d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1680,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=11;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1680,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.470000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-66627219b949ddbfaed8","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c289d5dba3c4838437e74881","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.08192600000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4574793e8c66e2a8536b6db977d83e5e27632db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1683,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=14;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1683,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.1926,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-189cc6eb852fb6af4039","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7ab36b3539a82712c5bb4fc","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0194","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a1329f16daa2ec564d1a422da7cebba1b85b5b0a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1703,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=34;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1703,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":1.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-925fd3c0c0c6f42b7ffd","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe6fd95a0dba3304cbd9ff21","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.11230000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a1329f16daa2ec564d1a422da7cebba1b85b5b0a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1679,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=10;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1679,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":11.23,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-63bf63d75edc1bd9a824","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b4b1852fba9bf5958d69022","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.1039","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ae488c82a330f49dbcc99a3ee07d2bfc15f7d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1681,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=12;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1681,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.39,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f34c7dfcc13e5d5f9499","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5c9240521729fd44def97079","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15960000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44986b19e66454506882ac17a3f8e45de25a1a1f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1675,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=6;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1675,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.960000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f8152059935ade1999f","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6a4c58205965c624f35b12b0","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2382","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c59b06bbf3cb80e681fc44751b0c49edb3394b65","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1671,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=2;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1671,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.82,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-81339b85b3a3b8a5693f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ccc2daba93c2478d2b11b2c4","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3712","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bb928b3b2dc262c012cf8ba1a4a22dc8d118ccf8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1714,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=45;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1714,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-enigma_eval_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c1d3889fe3488fe61e38","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-26fa9c6fc4d43e94aae55752","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0338","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4181560194d4637466fc9db0cf177e3582c37caa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1695,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"enigma_eval_external.csv:row=26;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1695,"metadata":{"epoch_file":"enigma_eval_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.38,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-de0db665b442f43bbd06","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6572ddc1e5c20a476e2822b7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f86c79dd0e51ed538c10bb442a47eae4a6307a23","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1930,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=223;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1930,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c3badb09e898a7252ee8","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8b2ec6dec50bda65f974b705","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa69ca45b8011586aef13532378842c79b5d3bb4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2100,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=489;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2100,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7dcd1cf729e513032083","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ceb6250906df6b68c6d42947","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_high","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa69ca45b8011586aef13532378842c79b5d3bb4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2277,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=843;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2277,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e79c202d71c10be91055","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-534f3d3a3230d86bbe35dba9","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa69ca45b8011586aef13532378842c79b5d3bb4","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2101,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=490;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2101,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5c4822fa86425d447ec2","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d695183ed177e92bfc15c03e","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa69ca45b8011586aef13532378842c79b5d3bb4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2248,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=800;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2248,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-818d51e448addd624b03","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e782cc15081f151db780967a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.78","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ff33218062d5118d6fa1cef162d1d751d59294a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2133,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=542;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2133,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.78,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-606617a9c1263c566436","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0beb50bd9675fad6f8b3d343","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae4ff1f99df28099e635574a1356cabee473a4a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2051,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=393;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2051,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-857c7bfbf967153e3ac1","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88930990717599be62b68967","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"147.76","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b7bbcc063e945cdce7560a9e90a1c37643ec716","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1740,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=26;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1740,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":147.76,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f0a4bb1644ebeb4dc722","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b7bee9d92a02e9ac2eaf8d8b","mappingCandidates":["deepseek/r1@0528"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek-R1-0528","modelRef":"chutes/DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f86c79dd0e51ed538c10bb442a47eae4a6307a23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2271,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=833;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2271,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4f2538d4ed503eb8fbdd","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53df77c27e301362df156951","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2113,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=513;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2113,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1f03c871b063eed29b8f","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f261f3995f874979fdf57dd","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1727,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=13;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1727,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b9e511333315b8c90373","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b31ab4da11ebf6e9a0a65f4","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1809,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=101;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1809,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-384137329eae930582c1","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0a2829f9464845edc6a0555","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1810,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=102;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1810,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-afca71cf9f26ef658325","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b9447d2da8ab000c16eebd3","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_medium","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2029,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=351;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2029,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b8c59f887ec7fca4288","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-949990cf45894d2c923c7899","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2090,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=471;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2090,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e0e9ea2bc33c931d45d8","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d4a6d1485050f8599e821238","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_xhigh","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"162.48","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"558b698b2b4d72f7af3f5c4621b99d2d682f24af","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1732,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=18;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1732,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":162.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ffe8eb43b0efb3bb1f96","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cf4f8b797a03d6b93cb81fb0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1739,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=25;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1739,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b74c920a2fc98d3c270e","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d13120a9d3d4b1a534a0dbb9","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_16K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2185,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=637;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2185,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c626cb08532300256311","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-57b0ce26b1c24ae5dc4a6e03","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_1K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2200,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=654;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2200,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-112977f805964f2e2165","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-02293efda7b024d61fd47b50","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1845,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=137;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1845,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc70285212170d70a931","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-61737660e0801984be6c67a7","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_8K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2198,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=651;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2198,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04e7863993ee7d407bff","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3e3ec8f8d75e420e4e336280","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_low","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2258,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=811;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2258,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e5f5a3c1c849479f8bbe","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b896cea46e31db88aa08a8cb","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"142.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54d9ff9829ce44c7ace48cc985a56b7fb18645cf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2050,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=392;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2050,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":142.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b8695f88f75a1f392ada","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6938b7eda5b2d8ff118ee30a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1907,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=200;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1907,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0dd5e7bb297029aa1831","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2e9decf668f771da5f8a8931","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_12K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2241,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=789;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2241,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-129efe24408557747199","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-46a3d8486175ed24c6eccc25","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1950,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=245;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1950,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fbdcdb68228eff92b9b7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-18b584a306ec03e013ee62bd","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_1K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2196,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=649;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2196,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc07e70e17a26217ad61","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-65b5dcae4d6d41cd66faee60","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_27K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1937,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=232;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1937,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4cee39f3085baf932343","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f30ba4b63a46b46a073d6dd7","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_2K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2255,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=808;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2255,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-879d60a90c12893babe6","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-315a46e328b49855c16da22d","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_32K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2105,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=495;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2105,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce75eb56df98daba61eb","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c885469151dd11b185f75335","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_8K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2192,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=644;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2192,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f5c9f9dc462db70debad","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-617621f23f4ff1a4d9549d5f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a57885e0b2e6ac9017ae2f7a6272a8e920f2da43","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2017,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=335;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2017,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c00700dbc907a752a8f9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bdd1fe10a480d906b2db8ae6","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1808,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=100;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1808,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-95d51855d3ae6db2b374","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d8140465a6b5a4878f531ea0","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_128K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1871,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=163;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1871,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c83f76279796d18ce666","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3b30834ffe56d7172f22c22","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1932,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=225;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1932,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cc2721f4dcd41be58365","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9d89c5b541388cd39aab91d6","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1886,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=179;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1886,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-039b0d1449b0cbde8eee","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-da22570db25682d30e7d8fe9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Epoch context-window suffix on the dated Claude Opus 4.5 release.","registry":"model_aliases.json"},"mappingNote":"Epoch context-window suffix on the dated Claude Opus 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_48K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1839,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=131;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1839,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d85beec310c85cc7193","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-afd17daecc732362a69e98de","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_64K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2172,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=623;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2172,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8dbe1f3c9f8997ab4e3d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f26b5885f1c577deaaf355d8","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_8K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2177,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=628;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2177,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-724e8d16ad491f9a39ce","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-550ead7548405150aa6311ea","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_high","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2250,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=803;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2250,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e0112f6ff1a8922e06c","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bcc6ca1eb8e8d53253515a8d","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_low","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2251,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=804;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2251,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6c78f951f69615357528","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5446b3f0eaff0ed6ac2cc315","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed2b9b7a1498997c2edb35ee72c61f268584d6f0","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2014,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=332;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2014,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-53b675f7f718210a3a12","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fcd7854f7e460065f9707a15","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1817,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=109;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1817,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-309e5a400213a9134019","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db27f9b616a690c75c80d505","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_120K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1899,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=192;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1899,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fe6a76b9d8264ceec03f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2bef93bac0b283389614d19","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1906,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=199;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1906,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-41b30cd9ca32a269b96a","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f2a0b8aade3cea69c49d277","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_64K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1905,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=198;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1905,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8ce07b751e3d5022a7f8","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8beae8bc03560ba16581cf10","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2011,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=329;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2011,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-08c2fa67e90d79e6c668","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6b154d392ffeb2b5b47bc0c5","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_low","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1818,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=110;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1818,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d399e29f939352e1946e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08eb78129c8efc565ba68d48","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1791,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=83;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1791,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1331d5f43d09ffa2db37","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-34f6eaf0477fe6cfafe583a3","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_medium","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2252,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=805;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2252,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-76a30e6911ebd215feca","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-40f3b31b29569dbb826ba8ae","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.98","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec2a98e7935cb93376e0e29e6eb54276a527a0a9","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2013,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=331;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2013,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-268ac41ea62eaedfa045","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9790b2fbe2d9e8fc1676ba6a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1816,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=108;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1816,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4db3b9e106cf4608f278","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-662fcb595906225774c4d132","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1722,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=8;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1722,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dcfec8233de4abff09fb","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-68cde21ff1525257f62b56e5","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_low","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1859,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=151;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1859,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-658dcda1e707bb1165a3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cf3b29521a9cc3c789dd2fe6","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1807,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=99;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1807,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-165d8895a172774b2391","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bdedd7a0836bccf3f011a73d","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_medium","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2023,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=345;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2023,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e3bc5cfeb26774256ba1","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ce7d795a3cdfe58d13f98867","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2095,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=476;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2095,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-477a297aea918c98577c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ea87d3749c793bc1ec49b209","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"234199be1184aa7e9d1fe937029cdba92a04bed6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1890,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=183;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1890,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-90e43c5b15d4699122e8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f87b0509e00525462865096d","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_24K","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2156,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=601;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2156,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6a5b3a1c105d37f380ac","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dda95f98f44d4349f98dcd3b","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1725,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=11;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1725,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-03ebb723f3124159d06a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fd27227ec5eb96acd128a9bd","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1806,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=98;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1806,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5447b7f375d463e67bbb","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-14ba5ee59b66f51c36c040f9","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1746,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=32;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1746,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0429bde3e7a0f4f4e3ad","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33b17f6df88ee059df117c0e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2022,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=344;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2022,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9fa658fc2d4068b815ff","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1cbae1b3b56a9e70444a976e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1805,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=97;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1805,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-edb204f146a67af66b2f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b90eaf89582840a9e9b82f9","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2010,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=328;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2010,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7af7a13ba03c7b41ebf6","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-47aa7c46d3bcc7846b21219b","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.73","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"777f32118167cf47351feb6484e2728741520747","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1836,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=128;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1836,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b1498013f1aaec222e99","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2e3f9878b268bab9e975502","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1804,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=96;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1804,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-36d886dacadd5a5e0756","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3c785c3b2b3734ecee711a7","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2027,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=349;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2027,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-70752fb7dd3b5a5d2707","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-766ce1d608af606c95c7028c","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1802,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=94;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1802,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5f58be67b251eb0c0a70","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-41b5d17bff39496f5669f8cf","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1731,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=17;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1731,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b4fafe44111c8f48d138","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5be46d4e95ccf47383aeb7a3","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_medium","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2031,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=353;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2031,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5dec111782ce11544c54","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6d505d1c00a1424df423759","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 5.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_unknown","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2089,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=470;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2089,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b45f55f97d848f096e49","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a5a39ba6c195aa2fbe58b0b9","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_xhigh","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf14a4b5535d0b23f59bebb4e4cae8ee33e822de","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2026,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=348;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2026,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f8c64bcc0fa74d551c3b","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-28df86e0cb659a1702211d8c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1734,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=20;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1734,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0389981462d91f4ee457","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a23eb17ce804037f984adcf8","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_12K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2243,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=792;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2243,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a70a578050c71c4df2e","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db1f7b183c6f4dbefd5807a6","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1934,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=227;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1934,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2107f7085eec5128f6dd","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd4ec918ca5c41fcba58c2ab","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_1K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2191,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=643;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2191,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a3d5ec9cf31e22e227b0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6eb7a1bbcfc4896095935a8c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_2K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2253,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=806;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2253,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d82f1aef9cb4adb17eb2","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c7c16e3b5da148501a8a174","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1883,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=176;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1883,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3e0bba746d7ab0709c32","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fdc92e6db7790c5c3a9cec94","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_48K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1837,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=129;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1837,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-62e79e41cb25558893b0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-787ff5af1c429a6081db60c6","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_59K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1925,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=218;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1925,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-926008b9f7fb50509a9b","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-92f26a9556a0447df3ca265c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_8K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2181,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=632;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2181,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f769f7ccc159b7589069","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6af39b1515ac3bc29f7fcb94","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_low","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2254,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=807;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2254,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":11,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-94bc8e876d78d4217dc8","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-12d1460273999d392df4dd97","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"647203824dfc269f55bd2bee123b91db5374c57b","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2015,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=333;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2015,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3a59cd5f35009460c211","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f10459a5bad426ec0679839","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1814,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=106;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1814,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ced933edbbad6111d36f","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-94f3de2856145d5c9d8c59a6","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_16K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1897,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=190;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1897,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-78edaeb2c3adc8647883","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-31fc3409ea0a6b1fd19cf59e","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_32K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1898,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=191;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1898,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e27f6abd07fe7027bb87","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-016556b11a3fe7ba8c2a4237","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1733,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=19;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1733,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e2ec6ecc3b32c7363bab","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e48cf957e16ca268ca5be186","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_low","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1815,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=107;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1815,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-70c66fe536e88f5685f1","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df2e40898442e44b608e4362","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1803,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=95;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1803,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5e42ea2eb85e9c34ae16","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-31281a414efbc10f0a4e0a56","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1863,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=155;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1863,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5d70d4f7e83e99e48986","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5acff9424685b7597b6f40a5","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4a5be510134b1f690caaa2505d962484f4b8a86","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2093,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=474;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2093,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5cfdad0a2262548a9184","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86b705da117ee468908eba99","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1821,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=113;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1821,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-642c98112ec2863bf4da","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-835a3976d85bfd3cbc73e524","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_16K","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2154,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=598;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2154,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bbe909229fb999f1fef6","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3fad886c767c1b8cba9da2c","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2036,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=358;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2036,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2c31d9775fa8803663c3","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ddb562303e0f80bb3c070f01","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_low","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2044,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=367;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2044,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5449d9c78cf1871b2539","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1104f0ccb849742b4e901261","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1801,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=93;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1801,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-13915d31976a8d692288","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-161a7106d1e9f9f979881f20","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_medium","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2040,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=363;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2040,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec408b7d30d5b474b785","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2af9313dac1869bde8f71ea7","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2012,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=330;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2012,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b5f6ea8123bc4c3024d","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-87ea3047a88fea11a4a5559b","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92dae0d15229a804b79cc3c1ca2e5c666b7da087","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1868,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=160;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1868,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f63e3087887b23a0722b","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8e178af3f6b28f21d9f54b44","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2","modelRef":"deepseek-chat","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a5c457f2750ce7e94c0c4700f2ebefad50cb067","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1846,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=138;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1846,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8474a3871792e781235c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a162232e5555b65b0feb9b9","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_high","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac60f454bb914c694cabd21ed3735ec9222dd9ba","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2168,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=619;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2168,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f524c2fe9568d6f7f4d","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7c4a0daa9e5cc4598e651681","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_low","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac60f454bb914c694cabd21ed3735ec9222dd9ba","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2170,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=621;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2170,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-110f0842345637196441","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd9bfef788b3b78fab4e79e4","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac60f454bb914c694cabd21ed3735ec9222dd9ba","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1825,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=117;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1825,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3b77e7174cd2038d8559","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f32d342d646ba39b57e56cb4","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_none","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac60f454bb914c694cabd21ed3735ec9222dd9ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2205,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=659;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2205,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3799a27493497c41277c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-00e17d90225bee0d47798af9","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_unknown","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac60f454bb914c694cabd21ed3735ec9222dd9ba","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2126,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=529;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2126,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c250c20de555cd0a5c86","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-00c5f400056dabaa1f41d8fb","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be76ee7f8bb1863eae07560b83a00a9bf375b17f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2167,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=618;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2167,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3205a9226a53c4f47c71","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64f5a921826103511135936c","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be76ee7f8bb1863eae07560b83a00a9bf375b17f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2165,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=616;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2165,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d287a6b15ff36affbe8a","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7224d21926a9f1e331192b9b","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be76ee7f8bb1863eae07560b83a00a9bf375b17f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1717,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=3;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1717,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-be64898e39a98122e7f6","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-691b4c67ff0019f0f86dc9e9","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_none","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be76ee7f8bb1863eae07560b83a00a9bf375b17f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2204,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=658;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2204,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5276be45df5df9ad69e5","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7cdb5411ecb89910647cb28","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_unknown","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be76ee7f8bb1863eae07560b83a00a9bf375b17f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2129,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=533;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2129,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d9eac8a79893e4427500","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8bc0e0b4574a3b29cf7d87cc","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07a79a64c510930ca17a5f453a325341d44e92bf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1790,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=82;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1790,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f36aba2d625ed84a2cec","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-94883d50bfcfaec0321b146f","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07a79a64c510930ca17a5f453a325341d44e92bf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1873,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=165;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1873,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-abbee058c734f1340c3d","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33f7be78c2d0b6b15f9132fe","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07a79a64c510930ca17a5f453a325341d44e92bf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1789,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=81;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1789,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f1e30c840bc5d49410a2","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1077f56f2c256f053cca051c","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_unknown","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07a79a64c510930ca17a5f453a325341d44e92bf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2161,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=610;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2161,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4045a79f88ba77d639b5","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3ad8abcb8ee824c22fca831","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a5c457f2750ce7e94c0c4700f2ebefad50cb067","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2178,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=629;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2178,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4587c444e07e435c559a","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-194362729390da3e6185b307","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of DeepSeek V3.2.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of DeepSeek V3.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"DeepSeek V3.2","modelRef":"fireworks/deepseek-v3p2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a5c457f2750ce7e94c0c4700f2ebefad50cb067","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1915,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=208;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1915,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59fce58ff56f45c493bf","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9922e0b7614c5d571822bd8b","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"148.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cbcd24108d4b31bb810a215acd7e7d053f5cf1a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1909,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=202;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1909,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":148.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6be840a368d9fa93cac8","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e78e9cb695af6309c8538d13","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"140.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e674620d09314268fad5bb4827327ea9ff4857b5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1916,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=209;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1916,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":140.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-09120e957bc2df8c1c36","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-229795892958f3668b3bc927","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.84","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dbbe9836d67e91314934e94c2257a24551e79564","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1880,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=173;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1880,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.84,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fd29472049753517563a","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a24767422d5a336b80cf1cc","mappingCandidates":["google/gemini-2.5-pro@2025-06-05"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro_32K","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.84","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dbbe9836d67e91314934e94c2257a24551e79564","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2186,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=638;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2186,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.84,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-01efc1927639b3d83a11","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b779a4768e08d327aa1d29aa","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.61","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b42b45ddc9eb970dd81efe0dc52a5c5c15c72ccc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1872,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=164;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1872,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.61,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e408201ff5e7baa92c4c","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-385bdc125efa0ff52c73b7df","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.61","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b42b45ddc9eb970dd81efe0dc52a5c5c15c72ccc","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1735,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=21;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1735,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.61,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8693bb6de5c063080e01","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9537be9618b23e64f313b9e9","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_low","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.61","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b42b45ddc9eb970dd81efe0dc52a5c5c15c72ccc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1820,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=112;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1820,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.61,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-592b9f7e1714433d7ca8","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4831e82a39bc69c3c8d56bb8","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_medium","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.61","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b42b45ddc9eb970dd81efe0dc52a5c5c15c72ccc","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1819,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=111;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1819,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.61,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-70ec844a78c100a75440","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac3e79ad233658dd64123529","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f443ba80da0cfa341a554cbf05479a22a8d27024","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2131,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=538;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2131,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-301f9bdd527cb041ef9b","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8ac9a97a63366310689dba6a","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f443ba80da0cfa341a554cbf05479a22a8d27024","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1795,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=87;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1795,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1aa8197aacc4eb695fe7","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-82f071026d0e8ac65ea466a4","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f443ba80da0cfa341a554cbf05479a22a8d27024","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1793,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=85;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1793,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-47bf13f2b7593fe0b62d","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c8ad7860b56084556597ef1","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f443ba80da0cfa341a554cbf05479a22a8d27024","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2189,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=641;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2189,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-34bc980a2b374cabf241","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-649459b2463d26556847beec","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f443ba80da0cfa341a554cbf05479a22a8d27024","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1794,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=86;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1794,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-24c1e1ad0a3a789be327","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9957be5b750136cb9f11af8d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92d402bd6cf775dccc71edbd4f8727efe7442b9f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1723,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=9;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1723,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1cc71776222d10591971","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b518868ea59ef9c85f6e701c","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_low","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92d402bd6cf775dccc71edbd4f8727efe7442b9f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1796,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=88;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1796,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-50d8a33ba302c546d501","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3198d67a82b82f0321c8214b","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_medium","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92d402bd6cf775dccc71edbd4f8727efe7442b9f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2054,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=396;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2054,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75989b028008c4deceda","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-44db85edb15428ed2c6c4def","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_minimal","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92d402bd6cf775dccc71edbd4f8727efe7442b9f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1857,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=149;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1857,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4bab812d5f9d5650fcaf","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fd14c590eb5c1254284843f3","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92d402bd6cf775dccc71edbd4f8727efe7442b9f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2025,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=347;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2025,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7ba94a98e70c0203e852","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2f722c53fbabd6eab7b16b79","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8622655ebafc09e47d6a90aeecc42de0e1288bac","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1831,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=123;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1831,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d97cfd66346b111965cb","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a495361730c4694d63b381fe","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8622655ebafc09e47d6a90aeecc42de0e1288bac","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1775,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=65;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1775,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0807e401de9a86746dca","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7dced62d9bca1b825c052316","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_medium","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8622655ebafc09e47d6a90aeecc42de0e1288bac","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2041,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=364;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2041,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f901854f1d53345527e5","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed96f02f2060784f7b0a04e9","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8622655ebafc09e47d6a90aeecc42de0e1288bac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1774,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=64;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1774,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ef0811b4f7a4c29e4782","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53b064837a716d72cb14dc40","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_unknown","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"154.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8622655ebafc09e47d6a90aeecc42de0e1288bac","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2124,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=527;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2124,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":154.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-278cb238283e81b6ea5e","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-40f8cfde24e0548a180be160","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c74dba5bb59f3df64aff687f422a710e75ac6857","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1720,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=6;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1720,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-28347c5deafe8bd14298","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-48575bf16bf6ed9fcdd30f95","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_low","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c74dba5bb59f3df64aff687f422a710e75ac6857","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2039,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=362;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2039,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-056fa534f779681303df","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be7fbf450949c7b9f99d269f","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_medium","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c74dba5bb59f3df64aff687f422a710e75ac6857","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2034,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=356;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2034,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c6170bd8c96d4e35d445","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-91aec8f7c77d5947a96fa5ae","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_unknown","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c74dba5bb59f3df64aff687f422a710e75ac6857","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2114,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=514;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2114,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bf3cb8444463378a5e7e","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-951b67e96a153f26954287a0","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.94","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bee095fe23bda3b077fcfb99c8c52fad1628409c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1913,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=206;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1913,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa61284423a6456210f9","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-764e64f0b1f5b994414a9f70","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbacaf66d6ff3bf2ab592e3c58314763f4f9cb71","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1900,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=193;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1900,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-53d9dff7d5715085f8f9","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f43bb6c136e4b720bf64bd35","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.85","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9771f51c989f95ed9122d85b3a1048587524f0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1742,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=28;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1742,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-19eb7c42fee4c1a9c264","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03cc1a2ffa8124cce1114191","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-5.2","modelRef":"glm-5.2_high","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ee1a26957a2ef13ca4d68e2eda455ba5b37af2a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2024,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=346;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2024,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d1397275e9582d8a8ff","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cf0d42e14db162ed76f894f9","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ee1a26957a2ef13ca4d68e2eda455ba5b37af2a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1869,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=161;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1869,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-00a6d2ad055e8336385c","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-976c348dadd55e2048f5134f","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ee1a26957a2ef13ca4d68e2eda455ba5b37af2a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1741,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=27;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1741,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-958ffbea3ee37c8636c0","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b4799dec78d2645764995694","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"152.02","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ee1a26957a2ef13ca4d68e2eda455ba5b37af2a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2097,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=479;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2097,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":152.02,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a4c040fb6c441b80615c","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2754bd83799b5a882b95577a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef8ff70ba3039d8a88cacb20694c1c78995625de","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1830,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=122;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1830,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1860cb3500755a381cb6","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-312824162ec934565564054c","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef8ff70ba3039d8a88cacb20694c1c78995625de","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1855,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=147;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1855,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-058100ddd29fb7d5b682","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77e46ff6e980897b68a1ed07","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef8ff70ba3039d8a88cacb20694c1c78995625de","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1854,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=146;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1854,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e086c0b1e7c92fa66995","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e92731b85c05b7eba1ca88ab","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef8ff70ba3039d8a88cacb20694c1c78995625de","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1843,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=135;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1843,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-90127b70584955789f1f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f2696d8c34e968940353b2ec","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef8ff70ba3039d8a88cacb20694c1c78995625de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2116,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=516;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2116,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c51f11bfd487017a0473","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1308e4b30101725121e5f503","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d08054bb0823498e1de3568daffd580bd882947d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1737,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=23;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1737,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dda25c949bfbdc01a7c2","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-033b33f65fba401d4cacf6c4","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d08054bb0823498e1de3568daffd580bd882947d","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1783,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=73;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1783,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-28d06400c7305bd24a7d","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f05d0fd96348e56859110b8","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d08054bb0823498e1de3568daffd580bd882947d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1910,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=203;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1910,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3736e3dc26c9d80dfcb7","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-feb3e6707a16e142cd4055c7","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d08054bb0823498e1de3568daffd580bd882947d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1782,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=72;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1782,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-15316c018ff2be607a11","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03634c95a94e48fb927ac9e3","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d08054bb0823498e1de3568daffd580bd882947d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2145,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=567;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2145,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9169a27e98360940cd6a","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-25c6759ab2c3b95105e3597a","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.11","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2543f8fd52b9d051e9a5f5b31ceb80b9e71e557b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1738,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=24;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1738,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d78897a25a8c6cdbb7c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1a795cc297867929acb494ff","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.11","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2543f8fd52b9d051e9a5f5b31ceb80b9e71e557b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1860,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=152;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1860,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-09376a624c0b03f1636d","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5afd4b8d6c7deba7cb8740b5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.11","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2543f8fd52b9d051e9a5f5b31ceb80b9e71e557b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1936,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=231;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1936,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-769f136203890b78e492","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f1b3161373d22a1ef9b26547","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.11","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2543f8fd52b9d051e9a5f5b31ceb80b9e71e557b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1861,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=153;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1861,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ec622678d85ebda11f0","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-074f9ab6e2bd7dc26ce06be9","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"141.11","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2543f8fd52b9d051e9a5f5b31ceb80b9e71e557b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2157,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=602;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2157,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":141.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c8e44118d9278fdf96e3","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c18b90098aa7dcc7835b278d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d8205135bced096d3bf26c7b311c4bc9c1c9ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1901,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=194;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1901,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce820cb996c581fd6701","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9a013bc9b49ed5cf2763f745","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d8205135bced096d3bf26c7b311c4bc9c1c9ab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1771,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=61;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1771,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-625138e2d0facb4a82a9","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-109d7559b0105544fd1bf706","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d8205135bced096d3bf26c7b311c4bc9c1c9ab","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1933,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=226;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1933,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d842bf0c255a3158256","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6814b710503208684f25f99d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d8205135bced096d3bf26c7b311c4bc9c1c9ab","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1763,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=49;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1763,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-126fb1dcb9807b4dcddf","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-988b75b125cc57b8fef23253","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d8205135bced096d3bf26c7b311c4bc9c1c9ab","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2016,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=334;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2016,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1604496b0b7d3156eb40","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a15d00f61f5b0e9ea39a4d89","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af057c03b530ef830324e9441395f02c7c9a6db9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1813,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=105;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1813,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ddc57334c741c85d22c","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98f5e4bdeea4ca183cbfcf2a","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af057c03b530ef830324e9441395f02c7c9a6db9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1921,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=214;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1921,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-661baa1208f19b834a66","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-164d59f31a145371ba468e96","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af057c03b530ef830324e9441395f02c7c9a6db9","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1920,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=213;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1920,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7dc4c3b30a0cc8901b41","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b4816dce4b2d8ba6d3556f46","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af057c03b530ef830324e9441395f02c7c9a6db9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1862,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=154;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1862,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d6fca540f5063292af38","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73c67f487d1f7377526e3a97","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af057c03b530ef830324e9441395f02c7c9a6db9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2115,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=515;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2115,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d12a840c7d6db28b0b1c","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5bfbff532544d85d4d158ac0","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af057c03b530ef830324e9441395f02c7c9a6db9","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1870,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=162;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1870,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b4f51ac746cf6a373ec3","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dac8c955062dd791005cb705","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a362d3c545a2736be87fc87b190993aa8576979","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2242,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=790;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2242,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-640d9f6522e2687abb84","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c1a37b89dc3b6761cfb2c08f","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11-webapp","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a362d3c545a2736be87fc87b190993aa8576979","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1914,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=207;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1914,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5157975787fb9b6acae1","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bfcd92a8d62b9c05ab754790","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a362d3c545a2736be87fc87b190993aa8576979","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2169,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=620;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2169,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9380057b65ee51aaac6b","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1e4ddadb27432833c8941a07","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a362d3c545a2736be87fc87b190993aa8576979","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2171,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=622;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2171,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0041b6275652f8cfd4ac","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bf5ed940690ea0cd18e97c7b","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"155.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a362d3c545a2736be87fc87b190993aa8576979","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1876,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=168;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1876,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":155.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7637b655dc4cf07c27a8","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a28b74335fc6ab298896377","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e58c39c5709bffbb47823283b6337df70d84629b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2215,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=737;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2215,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9bc3889be3b5b91ee745","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f0add221a2983dcffc39e017","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e58c39c5709bffbb47823283b6337df70d84629b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1895,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=188;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1895,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dc540a1b2550b5a4ea21","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-848e16c0c1f26f05afdfb061","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex_xhigh","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e58c39c5709bffbb47823283b6337df70d84629b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2211,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=679;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2211,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a6597811c9955cb5ebef","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6031377d83302d3ae9be9613","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dd5db0a3959c989d6f421a27044d3c1c13e9caf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1744,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=30;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1744,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-38b20d101f341a69cb58","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-39723867d286024924e22425","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dd5db0a3959c989d6f421a27044d3c1c13e9caf","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1852,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=144;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1852,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fb163bb756de16c3273b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-91c9bb1a7c0e1ed0522ccc67","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dd5db0a3959c989d6f421a27044d3c1c13e9caf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1850,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=142;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1850,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d2880e15494011ceb6e0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-65ee9a0dd096a130e5165370","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dd5db0a3959c989d6f421a27044d3c1c13e9caf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1851,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=143;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1851,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-edc5d03435e31520ee95","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9c058ef704c4efe12f50726a","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dd5db0a3959c989d6f421a27044d3c1c13e9caf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2094,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=475;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2094,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-71e26f5409cac62617e0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-21007b249146a135eec63dc0","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2dd5db0a3959c989d6f421a27044d3c1c13e9caf","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1840,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=132;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1840,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b385396514f65134423","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-58cef27096d529ef6b2f87f3","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e2e8c3c9437fc3029114075875dfa92ab637bbe","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1891,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=184;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1891,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8fead6222a089b29b033","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cb37a59807028f2997fe8a49","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e2e8c3c9437fc3029114075875dfa92ab637bbe","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2203,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=657;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2203,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bfc62383e545a664cda1","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bd670ef8f608990ef1ef9db1","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_medium","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e2e8c3c9437fc3029114075875dfa92ab637bbe","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2184,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=635;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2184,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f41c6a747e477b0893d","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-99ad9b1b757c50c3daf6005e","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e2e8c3c9437fc3029114075875dfa92ab637bbe","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1757,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=43;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1757,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-597b45053d2f02a79623","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a4ff48c6193bf8f4920079c7","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e2e8c3c9437fc3029114075875dfa92ab637bbe","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2125,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=528;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2125,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-37a66dc5ab2ee8cef3e6","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe9b863980d2e09fe6a2ad8a","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e2e8c3c9437fc3029114075875dfa92ab637bbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1750,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=36;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1750,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-94265913c8db1b75cd1f","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3c7735e78a677beed5f3186","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81d5c9d7b80e887b607eca046a0a137881924c1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1879,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=171;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1879,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d65d8f830f04be3f572b","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-842289ccd6034388d46d2839","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81d5c9d7b80e887b607eca046a0a137881924c1","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1787,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=79;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1787,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8bd7dac13ca24821aebd","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eb011ba8c488ce9aacb414b4","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_medium","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81d5c9d7b80e887b607eca046a0a137881924c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2188,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=640;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2188,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c13ef0be868defb653e6","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-70caf568f2fe5deb988ef922","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81d5c9d7b80e887b607eca046a0a137881924c1","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1785,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=77;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1785,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee4ef44eda7c78a778ed","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e737e30d04231b5e34bd90f5","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81d5c9d7b80e887b607eca046a0a137881924c1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2158,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=606;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2158,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-28d432e43d84aace97cf","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-000a8422ebe30d5bc4655c36","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"146.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81d5c9d7b80e887b607eca046a0a137881924c1","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2143,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=561;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2143,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":146.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e826e79c096ba3eb468b","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e245797371c34b85eb70a9c3","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05-web-app","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36fe592997080fa1cd09a9c79ad9623b87655d50","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1894,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=187;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1894,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f93da5d6c1ce43786525","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ffa4993dba3f40bca2dc46c3","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36fe592997080fa1cd09a9c79ad9623b87655d50","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2260,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=813;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2260,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f6ae3ebe8747c8158585","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f63c4638a09a94d0fe725a18","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36fe592997080fa1cd09a9c79ad9623b87655d50","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2233,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=775;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2233,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-de8c3a1a2b2c3afccfad","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f8bd1f30c8a7f43a94104b97","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36fe592997080fa1cd09a9c79ad9623b87655d50","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1875,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=167;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1875,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-11d0bda400a16a500188","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fbb55a6982c563d64fb63497","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.56","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1833,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=125;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1833,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.56,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f1245bc477e788690194","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b1d365c59d4e4bcec2b58837","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.71","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a125f7557b243b2fc3f2e1e40389a66ed166bd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2164,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=614;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2164,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.71,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca10199cf49e29b13b12","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c13d86a89aec33a747b1d8e","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix on the distinct GPT-5.5 Pro endpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix on the distinct GPT-5.5 Pro endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.71","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a125f7557b243b2fc3f2e1e40389a66ed166bd7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2238,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=786;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2238,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.71,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3568ee606c12d828a471","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a4be7fc098cac0729493a327","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.71","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a125f7557b243b2fc3f2e1e40389a66ed166bd7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1877,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=169;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1877,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.71,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad6ce2543cd49440a7b0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6213b0cbc0675b4579558b99","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1743,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=29;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1743,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-18ac20374482eb301554","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b4cea1fd68f2dfefdeae332","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1760,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=46;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1760,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-52bec9d4f9f33e5f67c6","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8afa9da2a03845ada32475d0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1724,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=10;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1724,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d7cc1194540b6ed8b45","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-14086d9f1ca828e03a294647","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1752,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=38;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1752,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1dc4cf7d3adee9a9ce9a","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-70ec95aa871356c603806df9","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2049,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=391;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2049,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-544d3ed8752e3535169f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1e19f6b8bc34538b7775a8e6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2ac9aa09e1fd021482ec8a42d4fd1df55d757ca","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1835,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=127;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1835,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d31cb28d7db485c62c32","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bea3075b63f5adbe458d1d88","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2037,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=359;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2037,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d8d3035f9284c032070","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9fe9b58bc1fa40a94959965","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1779,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=69;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1779,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d49ba86739cafc59ed42","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ddcfaaa0b1fc4e80b3c41a1","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1730,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=16;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1730,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-863ce66a7b9c59bc9666","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-083c6f97dca93ce11a038271","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2043,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=366;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2043,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-df58e78756f306df7030","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e47728c198608579be37f1a","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1767,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=53;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1767,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ffb46011bc2a60eef608","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e839ff2cde6969170fc198c8","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2123,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=526;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2123,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ae2f238c03880afc43f6","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-42c958f115641d26898944ec","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c384954881f966dd9057515bc5f37398b1086ca","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2035,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=357;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2035,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d1b6b90a21f8732880d7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e5660ebfd3257c5ed7578300","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1745,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=31;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1745,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a33b4cbb8d4627e95a32","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2748cb09278d4a18436a1757","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1758,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=44;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1758,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8f530e2225c2f2b2ec1d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f84522321833abf54a65a09","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1728,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=14;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1728,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f115a4594220e505cbda","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fd06215abed6837f80a684eb","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1726,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=12;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1726,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6c2514d19f72d075c4ef","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2302bdf1cb1cd49afde092fb","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1751,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=37;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1751,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7103ee2ddf38937f80d6","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-abee95fb8ce4467a9c473e54","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_promax","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1866,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=158;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1866,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cd31162160a2c96152c0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b97daff572237ae4496dd324","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_prounknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2240,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=788;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2240,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2c717191dd77bf68fcde","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a594f7d435a583c76cd0a899","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_proxhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2239,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=787;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2239,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e83de6d8db9d949fcab1","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b1fd36844ad75264a7011b61","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2092,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=473;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2092,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c5bd1b66acc70fbd6d97","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ec1d6fb517cc971b3131363","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"161.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f9019727a2274f72e0309b39eb5133fdf3c10a","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2030,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=352;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2030,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":161.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a254d3b4632b7a29b98","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ee35461cf56e754f64b51724","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2038,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=361;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2038,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f59060b1bd49b0f1396b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64a91034f90aa2bcbc6af4a6","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1765,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=51;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1765,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-74ae7586a95f42b07d6f","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cae40985cb38b7ba01bbb03b","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1729,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=15;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1729,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9aa176df645db5181044","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d94b270e5aeeeffd3742c7f7","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2042,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=365;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2042,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-67d399329694b5c453f6","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-169aa4757c06fd71daf2aef3","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1756,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=42;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1756,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2fc011c0700ba86f6b99","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c900fda7a412557394e6f6ba","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2122,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=525;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2122,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":7,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fadc83b812f341258999","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b89d3fa8192e18087af82bea","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"158.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3343993de492d3447a31131fb6628f54fda93b9c","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2033,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=355;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2033,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":158.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-09-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-60aa95edc277e961eba5","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d79b1475f73e11730760efe3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4 Fast","modelRef":"grok-4-fast","observedAt":"2025-09-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88bf52c5b29b3d6b774e312711dfc3d432759300","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2152,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=596;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2152,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2025-09-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c6264aad9a7a937a93e1","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-641ec2c1deb52fb813029e03","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00d4432e6d915d8b3e9c5d94a548a8631d78c979","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1867,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=159;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1867,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8a5bf69e6c2a4693f4a3","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e01196b60c5ce6db58240df8","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.5","modelRef":"grok-4.5_low","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00d4432e6d915d8b3e9c5d94a548a8631d78c979","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2173,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=624;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2173,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-205c07196a4dbef7671d","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-00b6c1c60ccaea9d25d59570","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.5","modelRef":"grok-4.5_medium","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00d4432e6d915d8b3e9c5d94a548a8631d78c979","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2166,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=617;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2166,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b3587b05adc014af69f9","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ecca46de11e5f0252d4ee605","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00d4432e6d915d8b3e9c5d94a548a8631d78c979","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2091,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=472;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2091,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cb894389edc11fcfea3b","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4b34b122f35ef8d139314632","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fed6310451884c7afec28544ad3e63bcb1579968","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1721,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=7;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1721,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a75a4366a0391d82874","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b4572601a3ee4d0451de126f","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.6","modelRef":"grok-4.6_low","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fed6310451884c7afec28544ad3e63bcb1579968","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2032,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=354;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2032,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d4d6a937c562b5c3ea5d","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6461f542a66253abefd0817d","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.6","modelRef":"grok-4.6_medium","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fed6310451884c7afec28544ad3e63bcb1579968","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2028,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=350;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2028,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6f7f5892ed0707ca9081","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a85506194a8d0611d1d30aac","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fed6310451884c7afec28544ad3e63bcb1579968","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2121,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=523;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2121,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-21753eae235d74b5e344","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a5d7811cd704ab0f593f9f91","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fed6310451884c7afec28544ad3e63bcb1579968","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1718,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=4;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1718,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-70bc91f9cca156d6fd09","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ebf8714929818e6edf2c1867","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"148.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cbcd24108d4b31bb810a215acd7e7d053f5cf1a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1902,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=195;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1902,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":148.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8427d0bb59e007d2a970","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1ddcfdb03070ca595f45993c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5_none","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"148.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cbcd24108d4b31bb810a215acd7e7d053f5cf1a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2237,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=782;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2237,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":148.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a1e01d552022b6805375","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5506526224a559718503c25c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"151.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e05443c2d13a8753b09c8e02b55d1f4f94a1d09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1736,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=22;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1736,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":151.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8fd3b5419be69e0c8387","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-040172d8bd9b6c1a099371bd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"150.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82e38057c14f27f0c53adfc7bcf400f19f6c7d8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1764,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=50;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1764,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":150.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ffe06ac30550b0bf6336","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-691c088f8550632adb60a80c","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55c2766211f161df5e31f598e252e81014b5b879","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1754,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=40;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1754,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e78bf03ccd09d21778fc","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03e21119daae018885dfdd41","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55c2766211f161df5e31f598e252e81014b5b879","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1753,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=39;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1753,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2304d2b568dce315339","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1621c1361994c2d148eb70b9","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55c2766211f161df5e31f598e252e81014b5b879","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1834,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=126;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1834,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-151918bcf821d730a577","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-461f4a160b7dd9584c5b1855","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"157.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55c2766211f161df5e31f598e252e81014b5b879","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2053,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=395;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2053,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":157.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-74474ba3af2c0c5285bf","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4686486e9d6bf89e429f5ee","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b062c0c3c011c79b72a85d54591a4fd1c7d744f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1786,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=78;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1786,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d93e4fe5005eb2a2a3b","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7d2440acd2fcb4964a0e52d6","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash_none","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.97","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b062c0c3c011c79b72a85d54591a4fd1c7d744f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1829,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=121;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1829,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-701fb0be56058663b999","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-acdfcdbbf6daa8bcd0bbaaae","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"147.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e484800a63004de527edd8d5c05ac54e195dbe96","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1772,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=62;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1772,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":147.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-84b473859ff00f0d6051","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2bd0b8586949b16051d84929","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus_none","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"147.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e484800a63004de527edd8d5c05ac54e195dbe96","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1827,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=119;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1827,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":147.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-47a0aafd9f925cfd824a","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed81d28be83f964d69f983d8","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5fce33f743ab2369f6537548e0407e8ca711db99","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1776,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=66;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1776,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.03,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9c808bdc401a17f396a6","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b3a0dea0567d22487098c6d4","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b_none","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"145.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5fce33f743ab2369f6537548e0407e8ca711db99","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1777,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=67;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1777,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":145.03,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e1648ba9a096f0b4ee8d","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d31bc8cc90c3fae43e5b7054","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3809439132c6f782560fb5e152964f72c72910d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1778,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=68;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1778,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-92579a5b6a8059fd6dbb","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ddf9ab96735a708f416abfd6","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash_none","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"144.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3809439132c6f782560fb5e152964f72c72910d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1828,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=120;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1828,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":144.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f492294aa4bd5b05a6e3","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-14199297801de9a20859a819","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4b5bd291dc3a30e804d54fc2de9c66d0f9eee49","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1759,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=45;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1759,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c7e317f737bdad05d80d","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-049a46ef675ebcd32c0917a9","mappingCandidates":["qwen/qwen3.6-max@preview"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on the explicitly preview Qwen3.6 Max release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on the explicitly preview Qwen3.6 Max release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview_none","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4b5bd291dc3a30e804d54fc2de9c66d0f9eee49","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1822,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=114;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1822,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b685e9eb758f610207ca","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7323e7a7eb59f773a7aa426","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4233b24de862ab89b033c83fe2dcd3c819853216","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1768,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=54;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1768,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.37,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-df851e45f0559cc81e3e","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7be294f4cd0f77a3f2986b4c","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus_none","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"149.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4233b24de862ab89b033c83fe2dcd3c819853216","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1826,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=118;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1826,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":149.37,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0d87bca85d30d5b7194f","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1fab0a2e28cf35eb1c5f93d3","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"153.55","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e16bcb86a7b71b95230a9a6b4a203a770e8a684","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1761,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=47;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1761,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":153.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-575b1d6d6299d5b1e31d","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-690cc054303473f00222f8fa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"148.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b3f497b696599e9901d3387ccfec523b297a734","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1780,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=70;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1780,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":148.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fe080c22131fdab35565","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-15c5d6f4e950cdf0c7ddf417","mappingCandidates":["qwen/qwen3.7-plus@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.7 Plus.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.7 Plus.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus_none","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"148.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b3f497b696599e9901d3387ccfec523b297a734","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1781,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=71;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1781,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":148.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1170b97e857952a14753","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20861e8b2481167afed52280","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.8 Max.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.8 Max.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_unknown","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.42","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9692933808cc42b14521efc56c36e40764d52859","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2128,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=532;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2128,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.42,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9de318df98d2ee9c5302","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e17aa49fd99483d38dc86464","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"156.42","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9692933808cc42b14521efc56c36e40764d52859","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1832,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=124;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1832,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":156.42,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eb7e1b1848ea2d13144a","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08c906b79845e4725cdf59de","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.94","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bee095fe23bda3b077fcfb99c8c52fad1628409c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1912,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=205;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":1912,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-epoch_capabilities_index","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-569eaeb07b3712748229","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2ed8bdc24d569240339ac31","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ECI Score","modelName":"GLM-4.7","modelRef":"zai-org/glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"143.94","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bee095fe23bda3b077fcfb99c8c52fad1628409c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2216,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"epoch_capabilities_index.csv:row=740;column=ECI Score","sourceMetricId":"ECI Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2216,"metadata":{"epoch_file":"epoch_capabilities_index.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"index","value":143.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b487477a7bd201cadc03","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-94fdc1cb2aa8198d70153e9f","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"DeepSeek V3.1","modelRef":"DeepSeek-V3.1","observedAt":"2025-08-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"58.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6c76c9d2ee53a2e1cafef3f0e2b037c49356c6c4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2342,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=59;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2342,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-08-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e7dacdb18425778601e","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-52544a2f15f1f2ed2a30fc94","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"60.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c0eb25afcb15a39974424975ba870f5c19895ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2311,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=28;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2311,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d5572d7bc28efb424b15","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-274db9774181bf8f8e888c06","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"537cb2874fcc53ac771724651b33d62eadde7b2e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2292,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=9;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2292,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0f5c66b89a5315f5d46","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c25447be7c8211badc13526","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6561110dbebef705ff674b238656a5683d178e04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2298,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=15;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2298,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-30820b0b3ff9bfb8bcbf","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c1c0e58c666de7cc65be3fa2","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"60.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"17ec6915da6b0407d55e9af6d051fa560eb09dcb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2304,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=21;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2304,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ebc3dfd219345c872a15","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6af0334e2219f01c2613823a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58d67e508e13bad5fc77aa1a8c6e06e7d6c2d22c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2315,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=32;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2315,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cc545f33361302847e9f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c35e9a4305363fd3778a15ab","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"820a20baacd0a351f4c59a506c8d6e56f613ddc5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2323,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=40;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2323,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f252f5509cd67ce70739","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ef9e42333e59a2802469b296","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_24K","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c6ff5969be1fb9a2a97bd807ac3ce2dd918276c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2327,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=44;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2327,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0fb5b140e0b3ac61f73c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2d7078eb2606bfc54618a2e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"58.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c6ff5969be1fb9a2a97bd807ac3ce2dd918276c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2338,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=55;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2338,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5f94f63cb8591428aaf6","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03d752f59cd304411bf41565","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4e8b2ef0d492c2293b780293a47aa46099e6977e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2287,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=4;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2287,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7d1d9ea42330b8e82da7","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-339edc1e0c416fd53bd6d8e6","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_16K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"75f0bdc2f6cddb1a156186557b634d41a27683c7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2297,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=14;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2297,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-06ebf4d85c7fa1101834","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-34dfc007df6cc3f9e3c4ffa7","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"75f0bdc2f6cddb1a156186557b634d41a27683c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2322,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=39;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2322,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d74f2dedfc4cbb1ff14f","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-deb215924378147d76f25ee4","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_16K","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"829db1c872aa6d188bceb310e6b25026a1e7bef9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2318,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=35;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2318,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d8b46fe2bd882633a58","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-673f6e3b2036959ccce72bab","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_unknown","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"55.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"acae41b2e5c722571504117ddff1235b71f7846e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2358,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=75;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2358,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f391b0e3fa3c0e23d905","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-def5b38df6ba8d344d2dfcf4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b7006b7b46d0bc485e57bee41a6bc9a3d504c94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2333,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=50;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2333,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-05165fc9cb89dcf10b05","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cdb2d06f803af57d32f34e69","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"60.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa786b659b75d3fb738615838bb542a0adc9c043","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2309,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=26;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2309,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77a7dc7bc0326616df99","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73a48f30cd53e219b9a8adb2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"830a5cdac816ac978082f153f33144f651fa228f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2332,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=49;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2332,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-da2afa72c91212d282ff","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1d21142b04c12c60174f289a","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"56.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"830a5cdac816ac978082f153f33144f651fa228f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2357,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=74;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2357,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e575a8888bd18675bb48","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4ea88e78d1851816f1f19f0a","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07a8242154e3e529cd4a8c57cab09abc6f795405","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2329,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=46;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2329,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1521aa0d42403a35801d","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d8af2b9dcc08b45fe4e1bb7d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"971c998425975fd5699e9c361e04e53496d7d2cb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2302,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=19;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2302,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c7b69541421c1e4632a1","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f91b63acd51e02289d90e91","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"583d2ddf11010d2d1448bab85a86cfb043110c4e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2294,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=11;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2294,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-788f9b2a79960bdcdfcb","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aee65c73e5d1532bcedbff22","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32d15036e1f209b9ee85de918775a2ef0ea128b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2301,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=18;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2301,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a7c040acfb79976c501","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1bd007720838575d9a59fc17","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0bcd183da0e5315b56658b5a00d0def3ec5b00bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2330,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=47;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2330,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ea2153649a46bc48f0d5","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b2b141ec60532a2b53c8a2ab","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"58.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5c0b7bfce57dd8f3b486c35624520d94cc88d12","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2341,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=58;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2341,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b0fe6382929446fa9123","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-303e20d684601e8f17fda3fb","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"60.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03608f05d8648e6af5b7da6f2de3058a7d3cb275","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2314,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=31;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2314,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8da5acc5f0a16874117f","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fc2764ff7b3044d647633bea","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"59.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3b936aa87b8216f8e515eab4fe80ee2ad162d7b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2324,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=41;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2324,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-39860e1e1813301bf855","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8fc84063e7cea42ae15db852","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"57.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a115616a615d8ada6ab8a4790497a7fec275ece","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2353,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=70;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2353,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d154165b7b6b949bfac2","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-107f30c05f50ec182b264628","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"57.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1361f13cdc1080ab5886da17f3bc6834439a92aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2349,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=66;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2349,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f1cfc6af9027baef7f5","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c479f2e794943f98dd6000e","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"61.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99159ebad4f7b0ee93c7926b8dd49e8d5a8372e3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2299,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=16;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2299,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-forecastbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32661d5f9ec66696b320","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e75e23da558ad5ecc97ce465","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Grok 4 Fast","modelRef":"grok-4-fast","observedAt":"2025-09-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"60.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"119e66c9d78ab5b285566283b3222b9ae382876d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2308,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"forecastbench_external.csv:row=25;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2308,"metadata":{"epoch_file":"forecastbench_external.csv","notes":null,"organization":"xAI","release_date":"2025-09-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6ee12b5e48b56a6620cd","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-43971de7cf8aa3bb79158f16","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.147","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b2fab3ec359b10aa27118c582e6a0a2c95e151de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2389,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=27;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2389,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.147,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9a25030e5d012c9e6327","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-ede78b68e6d3986d64b2222f","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5348","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34a953fccaa7b51cee6ad22416396707e44450b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2366,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=2;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2366,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.5348,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-71a1d2fb59226ce47eca","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-a255168995e2978dd05f71ae","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e536370c7dbdbc33f24d18621dd511e66e354530","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2383,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=20;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2383,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.2664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dbb51fe27d39d70fdd9d","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-38770ad6f0dbbe99265bc59a","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3854","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec1634b605f2d0b6873ea7aa40a855fa72e3a6dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2379,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=16;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2379,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.3854,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bccc5194de8f14d9621b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-c4fd62a18227766be17a7231","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.465","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0932790162851a8f7fa67947d2e73fbe6fe537e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2370,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=6;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2370,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c7349f2888db0dc7f56f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-dba0480e1e453076b2956421","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5338","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f06cf0ee6f9eb9ce3b4fb4970cff51c37440bfc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2367,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=3;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2367,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.5338,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-813af200ac7340a3d809","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-71446ab76bd9403b44e6c889","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.24310000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"788e0b1403921597cfe62f788019af1a03d2ea84","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2386,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=24;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2386,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.24310000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0cee3ba0a7de062827ce","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"claude-code","harnessId":"claude-code","httpStatus":200,"id":"pub-13dfc6bd48bb1b61a053988c","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.42730000000000007","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2cb1c0818ea9a299533c738e0dc57212dbf1ca52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2374,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=10;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2374,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.42730000000000007,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ef3ddc1e2094e0818273","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"chisel","harnessId":"chisel","httpStatus":200,"id":"pub-27bd4177d6211cd32e3fa5c2","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_unknown","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"chisel","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.188","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b333ad9d086506fb766815b1e3b0893b125fb5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2387,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=25;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2387,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"fraction","value":0.188,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f6492989b0c25ec91808","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-75ddfafbe26112a6829acd3a","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1764","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb2bc71c0a56aa1f3433e3faa558247aea802b60","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2388,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=26;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2388,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.1764,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e5edff548df4f435f9e9","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"chisel","harnessId":"chisel","httpStatus":200,"id":"pub-59efc1a08a0d92aa0bd378b9","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_unknown","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"chisel","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3437","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32f41d9d45d31e86be9314896e650834e0e4d808","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2380,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=17;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2380,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"fraction","value":0.3437,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-50603f9d7996c6db8d53","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"chisel","harnessId":"chisel","httpStatus":200,"id":"pub-17b0927f7185181fd2ae4fc3","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_unknown","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"chisel","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.43590000000000007","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"689bc6f7924589cbaa337f36b84098126c395884","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2372,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=8;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2372,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"fraction","value":0.43590000000000007,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-537ef61f560c12c8bfc8","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-8cba59f43c139672c245517f","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.245","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf8bd0f2ccb10d6c9a3ff69b74ff7ce954f2905e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2385,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=23;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2385,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.245,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-53d14b517b7657b41d58","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-4073bdbe6133faf507f4739e","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_unknown","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2704","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ed1050419639ed82e2e8af66af82526d0ac725e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2382,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=19;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2382,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.2704,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e6eb5b973102777fb7e5","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-c5f0edbf4854db1b2de86f87","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.42960000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"844331d354496fb000f33cb4b38ce5e781b5976b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2373,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=9;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2373,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.42960000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-903ddecf4ba51d0af159","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-b4fdf313cc2db63efb2f40bc","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3981","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc291c344bca79a44f36d3ff71d099b2fe1270da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2378,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=15;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2378,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.3981,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1ae1be293791e4535fc5","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-0a5af6fe9473f47264a12443","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47490000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a53599e8210cb03acf4c7bacf058e737a10b212","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2369,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=5;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2369,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.47490000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9c564240d2bc75ea215a","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"codex","harnessId":"codex","httpStatus":200,"id":"pub-019a1a70a84e06cedc9a7e0d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4131","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6531de8366eb102318b4d91178e6eeef81ea33e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2377,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=14;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2377,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.4131,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f29e9e29198c7eb9c5f5","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"grok-build","harnessId":"grok-build","httpStatus":200,"id":"pub-4f82871178a1d984bcbffdfc","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"grok-build","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4244","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b8547d10a60dbc0f318e41924afe1d42b1eac6d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2375,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=11;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2375,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"fraction","value":0.4244,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec18c0938c49d6715588","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"grok-build","harnessId":"grok-build","httpStatus":200,"id":"pub-fed89667d781afe4c3e2a1f6","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"grok-build","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4801000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4468c36fcbf0639d5fd8bc22f2869109da6fa0a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2368,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=4;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2368,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"fraction","value":0.4801000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bb295500c6583367c772","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-0565b8b057a1a98cb94bf837","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3006","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8114f74689c116a4d8c8ee8cfdc5924f810ef77f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2381,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=18;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2381,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.3006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a9e0c35f02d266aca467","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-14a7818eee019bc453c24ea6","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.44170000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"971f25a2f9060bace9666848937e05466e76660b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2371,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=7;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2371,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.44170000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiercode_external","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0725c36d2ce79ae49355","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"mini-swe-agent","harnessId":"mini-swe-agent","httpStatus":200,"id":"pub-1ece6d73862a47a13afb88d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Main score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"mini-swe-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.102","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a4708bdd409e2b1f2e1bacce4b72d985848b2ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2391,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiercode_external.csv:row=31;column=Main score","sourceMetricId":"Main score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2391,"metadata":{"epoch_file":"frontiercode_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:kind=coding-agent"],"subjectType":"system","unit":"fraction","value":0.102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-efdc789a67615a550add","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bf9ec78ba72bb8b552484f94","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.020833333333333332","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7632274ddf3a4064fe4d16505ff6fbd5abbba21","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2543,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=50;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2543,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.083333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a58dfabd5c297ed84aec","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b699fa684399df707bb9d620","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e1e2817f4b8a15a5d0f3062b92cd3a00d15fe4b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2557,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=64;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2557,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-010aa0c5729139d5a584","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-23ea53b1e970a0fa3b929bd2","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_27K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e1e2817f4b8a15a5d0f3062b92cd3a00d15fe4b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2561,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=68;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2561,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7048e51c408e7269808d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-50de2511ddcd92a0eac85221","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5bdf2f709453dc8452592c0718d61c9d665b328","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2534,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2534,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f8d8e66d20acc1378df9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32d74dc301d46379286616c4","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.020833333333333332","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5bdf2f709453dc8452592c0718d61c9d665b328","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2535,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2535,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.083333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bfbf44638159e21bf6b0","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-58d0571abb10d0dcfee32fdb","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5bdf2f709453dc8452592c0718d61c9d665b328","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2536,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=43;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2536,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8ee89b2b0676eda3b3a2","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa444b58d4e89d2f4a51b9ac","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.14583333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57d63b9c55e6b08fe0ae5b46ab97a9d6aa5d4b81","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2520,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=27;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2520,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":14.583333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-094f750372ecf6b633c1","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5033eb55a0e1a1b047d9b49b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.20833333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57d63b9c55e6b08fe0ae5b46ab97a9d6aa5d4b81","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2519,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=26;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2519,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":20.833333333333336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9f44e4d2fbdb78b119ec","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f16e01092c306de50ba5d607","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_64K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.20833333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57d63b9c55e6b08fe0ae5b46ab97a9d6aa5d4b81","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2518,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2518,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":20.833333333333336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ba86d24a3472a3aecf6b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5283c6443ebc280abbc86acb","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.229","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57d63b9c55e6b08fe0ae5b46ab97a9d6aa5d4b81","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2517,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=24;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2517,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":22.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-efc5093e47e1db852d06","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86c1be7e4ca51b4e286567a3","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.22916666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3de47aebdd4243bf4ad7ded20aa28868b601062","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2508,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2508,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":22.916666666666664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-725dc64431412a9a23fa","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fa721ddfff8c6b342d0958fb","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76e24455bfaa8d1cd97b9f2483f593b1b812021e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2495,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2495,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":31.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ac944a7facb324e2f579","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-857015e57c1d8b6e79d8891c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.020833333333333332","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21486d8be4c39c9b7e65ff648d3755346e3b3d2d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2547,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=54;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2547,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.083333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9043f76ae28c89c97ef8","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5622074249600d7ecbc24d41","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21486d8be4c39c9b7e65ff648d3755346e3b3d2d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2544,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2544,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b013d6ea4e43bc4c1352","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-370a6ccbd3c3ff9c27a206e8","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_16K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.083","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f931c96b3ff3a799f72c9949402cc7d5df1e5f8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2514,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=21;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2514,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":8.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-01","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-11e9e0fee2bcedd7cc0f","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f6e63e510b369e1f1d98c367","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of DeepSeek V3.2.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of DeepSeek V3.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V3.2","modelRef":"fireworks/deepseek-v3p2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.021","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4e00581d47969cce018b4a61f1c2285c33862ad3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2526,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2526,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-94475de6ea63ef309eba","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-06596c038e135ec040b8a931","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.042","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb018208dcaab0178ac26eb81c83119949790053","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2521,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2521,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-05f1cd169fe37f9715b6","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-176794f8d17194261f1db20e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bf61b96d04c5ea23e9395e977412d7b785ec306","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2524,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2524,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ac765852f0a55e3464c7","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-510c760aaf4760741dfbf381","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdd6a35959940ddeb86128c88d6ce0cf13d415bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2554,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=61;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2554,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-14ffbb6f3bd46f95779a","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2a8f3d25796be67ee19bb217","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.167","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b100043b93eb4fc1e86b69dbfb6a0831f3c4d4d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2515,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=22;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2515,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":16.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7691b85375dcf101c64f","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bc0a4ad75cd6e3481327121a","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.14583333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78e84ac1604c249a84720c77e2d9eae1b61d5d22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2497,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2497,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":14.583333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-11","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ce4c2d09dce6983de879","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-71725b2a2e43ef7847db3988","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.021","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d9e649ee83cd6078df3fa6e3d3f8bf862fb23ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2516,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=23;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2516,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-04-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8e27ddaa2c1175279092","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d1a3b8949416b84aa941838a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85acff7921ffba940095cd39ab093667e8141d35","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2502,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2502,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":12.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-af4c2fc81374e6ff348f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4fa5acde1ab818d29a2101c3","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"965e4772b76ac772fcaa929c572ac393361fdda6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2542,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=49;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2542,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":12.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bc947c9c4e85c5912f52","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ab2542ffc397ee870f212d51","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"965e4772b76ac772fcaa929c572ac393361fdda6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2550,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=57;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2550,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":6.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b9596096fd6e10b7a94d","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b862580c9bf1e8a07ab517f9","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30b0342b86873dd380a210b23edfd6f7364f717e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2540,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=47;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2540,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":6.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-56b9da6fd6dfcf0086a5","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7be15e375bfa3163e1e98c09","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30b0342b86873dd380a210b23edfd6f7364f717e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2551,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=58;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2551,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-98f43a323e2c151030d1","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86325ed9b5720e9450b7cc28","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"accb94eac3a8c14cc24703af122f857244fb5f5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2541,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=48;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2541,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7cdfa00e78dce3050553","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4d53b9c22530963c33dbab43","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.020833333333333332","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"accb94eac3a8c14cc24703af122f857244fb5f5f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2552,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=59;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2552,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.083333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bcc2704a4441f46a5e39","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7c7935c6a4970d82f9686a90","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"31bbd18e3983047ab157c3ad558b4fda29f19b7d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2539,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=46;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2539,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":12.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7bbfe589479d99aae2f4","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b84818dda52520d0c602d63e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"31bbd18e3983047ab157c3ad558b4fda29f19b7d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2538,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=45;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2538,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7799955382a17b011223","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eae00036a727ec55a1730f4e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.188","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffd544d5965b18c42e456e5e4fd6d98a8f0c5fbb","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2528,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2528,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":18.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-70dfe188edb232b67cfa","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-78484996daff0b388a5e0110","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffd544d5965b18c42e456e5e4fd6d98a8f0c5fbb","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2530,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=37;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2530,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":6.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e3e743a16ea99c654d1f","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-efe8e364ecf07029332eb18e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.167","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffd544d5965b18c42e456e5e4fd6d98a8f0c5fbb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2529,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2529,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":16.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4a2dac5ab4f3bfe97ed0","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ff80999833d12d48f294b492","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.188","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffd544d5965b18c42e456e5e4fd6d98a8f0c5fbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2527,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2527,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":18.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8665a33294d9d0203682","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-056b72e9b6a36bb51337511c","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11-webapp","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.313","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3e3a039e662a62c24230a84583b9d0a4eea3076","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2523,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2523,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":31.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9d8e5dc7bf5a692e34ae","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7edcfdc4be5bb1d6c7a401b","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.271","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a7c73ee7f9ec49ef520210b975897dd7b81e8f10","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2512,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2512,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":27.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-74ea197a7bd35f4a8530","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8ff5ecf8fd3b27cddab80391","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0208","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7953e3a3b6d49d9288a4b01f9fa74411daee336","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2509,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=16;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2509,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.08,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a1071028d9faa3779466","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-717db7da6dc77b912dd1abda","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4755bab23b868b687555cc93c5a03e760d687fee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2510,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2510,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":6.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fd4589e749d47943af6b","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b42a7e920a5d178a194c4d1","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05-web-app","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc971ce3fc62013a3d816a68e57c819469d14303","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2513,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2513,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":37.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7f7f75734678564d5f35","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c63c80020ca9f0a0de80dfa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.1458","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c5591892341e0a6798b7bcc744c519528d74986","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2504,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2504,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":14.580000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-25","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c0a7ab81191c4f647589","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-374d24ef39057e16b72daa79","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8285f6ac08fe63fe4f08f65e599f0ac35d79e73c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2501,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2501,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-02-16","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8fbb893639a97b3024b8","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0cee971842160c3829d424ce","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.020833333333333332","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71519e8a7ecb1ad5796a083632afee9a08844b1f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2498,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=5;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2498,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":2.083333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-04-27","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2a21db629ccb3c39169a","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f8a9a027f13e5e5d70ad258c","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a379fb19fdc7c19c4d9cfa0976d7bbfbc192001","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2500,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2500,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9b0131a43c1d91cc85f9","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-06fee06708b8cbac242878f1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041666666666666664","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"889f576db3373d2f36a56b4b54a19bb42868a05d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2496,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2496,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":4.166666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2026-03-31","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bdf5e6fda332b77805c3","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-97e2658a0b2ffeea60368c4b","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.08333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72e1c8bf6a1550053f247d91fcb0220850c31b1f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2499,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2499,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":8.333333333333332,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontiermath_tier_4","benchmarkName":"FrontierMath Tier 4 v2 (reported snapshot)","benchmarkVersion":"2025-12-22","benchmarkVersionHint":"frontiermath@v2-tier-4","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-64c41c52775942f2f4d9","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-de572d1b53f11e1ba91c90ce","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9420ffe10c3ddb4c7d349c39fd2fe8bf58a8ebf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2522,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath_tier_4.csv:row=29;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2522,"metadata":{"epoch_file":"frontiermath_tier_4.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0a181dcc36d3ce949aa","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-416d843faa010022166c3c63","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"2.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7ee64ee1ce6c6a8d346c7dd288c66686d104b209","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2574,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=9;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2574,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":2.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-60b9263b16e28302bf88","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-9c901a6c379d755355d96055","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"9.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d93ea04fbfa2be39f1da0588f3cb28192327c9dd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2569,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=4;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2569,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":9.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e354bc755e14fadd2a92","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-9b34a32fc0faef6e1553c176","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"7.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"815bd9488cebb9311e16c2df1565545037b27d6e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2568,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=3;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2568,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":7.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-690efd00331feefcb605","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-cf972abb1ff992ac30cf8b2b","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"5.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4faa41f19981221606f3c9537d49a7a6ce668ffa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2575,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=10;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2575,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":5.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-247c4754975b2c6b2012","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-994e05ac310588b5ba638f7e","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_unknown","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"13.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81c4a8f0b7e1e1beeab90d100ee0495714d4bd04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2577,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=12;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2577,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d4fee3d36ca1a422a0e6","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Gemini CLI","harnessId":"Gemini CLI","httpStatus":200,"id":"pub-85e3190298991ebc9cd53745","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Gemini CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"13.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cffa8cbad5dee14b12b4f622e53150b51524cd5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2571,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=6;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2571,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":13.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8b8bbcfb9995e31195a9","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-885648dc9f5a2a1148046722","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"12.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7166a70104403991f42e62ddaae2dd6be4494ad0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2576,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=11;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2576,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":12.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-63b1d6518bc817b9f256","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-feb6a2103f712176e44dc37a","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"6.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5aec07a974f495d349cb88720c2a95f29e71789","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2580,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=15;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2580,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":6.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ed0f113a8f2691d0992a","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-4ae7fa2f783866dc0c374096","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"GLM-5.3","modelRef":"glm-5.3_unknown","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"4.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f0b4d4741b1660baf670735a5b95e2dd57e797ef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2583,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=18;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2583,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c0a53ac443201d352105","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex","harnessId":"Codex","httpStatus":200,"id":"pub-adb74b255bba76a9b863cb66","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ddc4ac9fd90d010efb29f3e2354da4e02d2bcef9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2570,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=5;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2570,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex"],"subjectType":"system","unit":"rank","value":8.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-973b8699ca02900d58da","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex","harnessId":"Codex","httpStatus":200,"id":"pub-d2d54b0d15331230f8451870","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f25a67e792408ec9c188233a221de3a1f13bcf0f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2567,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=2;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2567,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex"],"subjectType":"system","unit":"rank","value":8.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3a3ebe6a9f97daff4fae","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Grok CLI","harnessId":"Grok CLI","httpStatus":200,"id":"pub-aa8dfebcc89f0af203caeb87","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Grok CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"6.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f315e07392abb6232889778faa0388bb18c98f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2579,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=14;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2579,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":6.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9187d25973c3e94ee7ac","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Grok CLI","harnessId":"Grok CLI","httpStatus":200,"id":"pub-26c88ce32ae61bb135be2db0","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Grok CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"3.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"189bf4d8ef6a8fbfd99c6fd67b77bc1989ed4700","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2582,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=17;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2582,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":3.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8381d350e375112dcbc9","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Kimi CLI","harnessId":"Kimi CLI","httpStatus":200,"id":"pub-954e13ab2034ecb61db3eade","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Kimi CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"14.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"acb27b4c6bb367a2649ef530dc24f2ed0ef55fa7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2573,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=8;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2573,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":14.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5ca2d2a906e369c642c9","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Kimi CLI","harnessId":"Kimi CLI","httpStatus":200,"id":"pub-42feab8cc12bf7b0a38d35e4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Kimi CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"11.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5095ba06ca6adfc879ca881e29db43b6eff6b0e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2572,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=7;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2572,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":11.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-frontierswe_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9cf6a5f2f3e380955bac","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Qwen Code","harnessId":"Qwen Code","httpStatus":200,"id":"pub-c2cd58f13c57b8e6037f0f75","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Implementation rank","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Qwen Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"15.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"496ba42099753fb4a774a93aea3d3acd3f32efbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2578,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontierswe_external.csv:row=13;column=Implementation rank","sourceMetricId":"Implementation rank","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2578,"metadata":{"epoch_file":"frontierswe_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"rank","value":15.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d853b695aa720c93da8","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a4093c93b7c9955b8902ec57","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41b0c3313bb45b3794f58e69a6403dbc6b1299a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2605,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=23;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2605,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a815bd6f235d6e1ef9ed","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-17120088921f7ff19a25db88","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.009136320401503086","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69ee88448e5aed3c728a81911a1d4ef6401a9abe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2599,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=17;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2599,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.9136320401503086,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1126792a4ff71e0ac417","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-328dc4e58e97d74acb7ec1d7","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7447198828312318","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf86cdec72edd2849a482dd59edaeea3679a4362","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2585,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=3;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2585,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.47198828312318,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c6c350b6cfe3d075f316","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f1ee0cca3575f6e05d54cb2d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4411559216788414","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bb80d953baa0a02205257808764404958dc6888b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2594,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=12;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2594,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.11559216788414,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6bf97ae4d62dc37b94b0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-092eabfcf579ee6005e1c91b","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43811361876085897","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ded09607a70efc7c281a091fa6f474987722127","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2595,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=13;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2595,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.8113618760859,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0feabc09d9633cc94ef6","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e163c80d65a4a287bf541a5a","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7090450062635527","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d34051cd04cc1be6f56ba3b69f3bcc72c1cf7a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2586,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=4;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2586,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":70.90450062635527,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2c6c8659ad453d498b1c","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32839b702caf926f78a27060","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 5.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_unknown","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7959852171423216","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"415e0fac2019a30837fe5c89301cbdfde90f0768","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2584,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=2;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2584,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":79.59852171423216,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b093cd3c59f97e5157e","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5263608d9e3a0cda8e267d35","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4875663483204475","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f9967a9b39efd4c36e63cf9aa792442fece05a4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2591,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=9;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2591,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.756634832044746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1415059d90a040653313","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88f1fa6ead8cc2dce32241ef","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6526299757455419","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4252f34d65e2d3f596c4a102cab5d3d6e7b1cd11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2588,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=6;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2588,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":65.26299757455419,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-303c7f35ca32c0378f71","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c0889ed8dc2ff63004b8ddef","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.008462790979968672","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ebf3f9613d776c85ccf676b5a5aa05592d17d445","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2602,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=20;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2602,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.8462790979968672,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f9b88f2101b1249bba94","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33f2e9218b439dca5a81895f","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.06701366402557545","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0097361ec8938a8e39f450ea4847780a4705a1be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2597,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=15;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2597,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.701366402557545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a69addf306d2ebd7c41a","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6f23e45c240d48f41019108f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8ed0370824439e03d853b0f42e7df0f52a05eb1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2606,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=24;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2606,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-794fd8984f9ff1def8df","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4bdae6c21934cd1fcfe2ea5e","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.00017218231881747017","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8aa22a6a0bd15cd5c8970c19acb452ae49f8fbc4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2604,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=22;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2604,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.017218231881747016,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-978d56c631fbc85b4642","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac852cd6115d100729a67455","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4514817826662626","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e95f177eac2a492f78393c67d9b526885ef20618","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2593,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=11;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2593,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.14817826662626,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-54320f6ac95913c7b811","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d08a25db7250e97db320d3d","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5322037178151383","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3ceab29f4e543c4681aa77fac27c57ddb9b1f7ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2589,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=7;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2589,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.22037178151383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-568afc55437f7e2b373b","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a09301f76a459c867e65f984","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5259227864005247","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"38f8324b9c40d498b43904ab30d19b9a971fdbe5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2590,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=8;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2590,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":52.59227864005247,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a4e2d3e8af6bc3b909bf","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ab062fe6ce8a2a400fb7430","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6541267565853209","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"79f939d757c0ae5a52cbbd2238b0f41b0277bf34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2587,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=5;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2587,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":65.4126756585321,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-568207eafedc92611e4c","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-67fc7118b9dfb22bfe985f84","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.008624932115315736","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa27154071924ac03a9d6daa9a3bfdcfcbb68177","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2600,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=18;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2600,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.8624932115315737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8197a525f34ae1a10d75","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cb0b5fed1b936089a4655db9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.008526563108968812","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"629e654781021580c4259660b16f356887e67f4a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2601,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=19;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2601,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.8526563108968812,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f7a17279ae62aa1b428f","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6f49418057636b26f67a69d","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.48346855379537595","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92f4136ce656d0507b5f1ca3f68b4f168c44bd38","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2592,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=10;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2592,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.346855379537594,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gbaeval_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3e7030ba14b1cf81136a","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba7cc002c9e522b00d5f164a","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.004066608893369503","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9913fb5f99b89cb5cfa56b7488991ee89114f93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2603,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gbaeval_external.csv:row=21;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2603,"metadata":{"epoch_file":"gbaeval_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":0.40666088933695027,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d1b7e3df48e903b0cf7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-11dcce2aa0c91b199af89013","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.298","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52a0ca3e0822f9ba2d964d1b7eaab4031965aaa5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2617,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=15;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2617,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ccd110a74dd6f4003f83","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/leaderboards/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03342d53cf4475446b59c399","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52a0ca3e0822f9ba2d964d1b7eaab4031965aaa5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2607,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=2;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2607,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":"https://surgehq.ai/leaderboards/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c517b0b18b5caab3cccc","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-492423a37a5eb96bf1cdfde1","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5fd7002a398a6d6c60543caaf808b83fb60fdd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2610,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=5;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2610,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ede228bcb9829233a865","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f02b3aac76fe2b72746b3609","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ea71165e8a7b3453d0bd0c855274e6a9b6af6c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2609,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=4;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2609,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-948e6478604419e9cb59","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4f3b659462161f9fcaa5e413","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53f5b8687098a4066a6eee37fac4d2b8cbce1823","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2619,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=17;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2619,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0bb5a38d80aa3dc5901","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa33769524c8e2111e8665f5","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"601ec439298c698cf5ecf93fcc4a5f4f1c18cd2b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2622,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=20;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2622,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e65eff58238aa81247df","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-938b3c722a1a14e5d49a0c29","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ecc9ad299695486643caed9a76b16f1ed37601ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2611,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=6;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2611,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f7860cf7cccba3b4f7f4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2422d85f6774468b02890de5","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76734fdf6d4e7511d14a9cfeae56625eb586b0ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2626,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=24;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2626,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1cdc109fc7154aae3259","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2fe1709d8f1275c321e9986","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eae9af0398fa39a7a7a30922207d2bfe3722fdeb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2612,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=7;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2612,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-885c8887338ba6b51341","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-856514968011f1e8510bb2a2","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f9133b87c200e4d60cbdc73847e2cb74c8104db","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2629,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=27;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2629,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-41215384a7570f7be94c","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9d3017f89c8fcd88fcea459","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_unknown","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f9133b87c200e4d60cbdc73847e2cb74c8104db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2624,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=22;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2624,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6364af08ebb04b8e8319","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7e89aebb49c5f98c0d599577","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"205cdbb78faf7c3487d1e98bbcb1425a9fb047d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2608,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=3;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2608,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.26,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e276413c97f467a62b2","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be2d8d70cf988dadeaf4aaec","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.227","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c29fb146b62b6b81ce28394891c041e57957f893","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2620,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=18;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2620,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.227,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4ca18ec24ba2bfad21c2","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1673318876635b60705e79c8","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.307","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc48ae6cd65e9c1b24308628d989cf7611153b9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2628,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=26;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2628,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.307,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-91fa0c14921c314834c3","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-503fe2d72e97ff575b31b407","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.307","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc48ae6cd65e9c1b24308628d989cf7611153b9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2616,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=14;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2616,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.307,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a092631bec301893c911","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6c218caef7dbb933fd2724e4","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.247","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fbdec3b93e50b0d9a34f6e78efb0139100df59ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2618,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=16;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2618,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.247,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-86f091fa3025a15a6754","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe3e9d021318b56ee0321318","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"088471b98b2a26bec197551906bfd9fb547dd799","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2625,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=23;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2625,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8faba541d5b031c79233","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a900f736d40a157798ed4164","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"06/06/2026","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c84ee013779f969cd3539f3582a15907f700b7b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2613,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=8;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2613,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdp_pdf_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-365e0bb2b708c2d39c61","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://surgehq.ai/benchmarks/gdp-pdf","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-af83131a9b6b752591fbdb0f","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"GDP.pdf score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ec862f0c9e6b2b388576a8d31a7fa8f972f91d0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2621,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdp_pdf_external.csv:row=19;column=GDP.pdf score","sourceMetricId":"GDP.pdf score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2621,"metadata":{"epoch_file":"gdp_pdf_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":"https://surgehq.ai/benchmarks/gdp-pdf"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdpval_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a1a0c8beacf47f4e48e3","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c001a9fd7505d46d0d68fdee","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Win Rate (%)","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.455","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66b48e530a30eb49455eb858a0e1432aa4582c09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2634,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdpval_external.csv:row=3;column=Win Rate (%)","sourceMetricId":"Win Rate (%)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2634,"metadata":{"epoch_file":"gdpval_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdpval_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-db7172e72a758c6286dc","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d9f6a6d64f8f3d65ebb6ca25","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Win Rate (%)","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.425","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87cf03ab53921ab29df15a9ab0a544cf69f6f274","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2636,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdpval_external.csv:row=5;column=Win Rate (%)","sourceMetricId":"Win Rate (%)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2636,"metadata":{"epoch_file":"gdpval_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdpval_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-909951a31415ceba6312","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5526125b30708640b9afabc3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Win Rate (%)","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.233","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b2ecf7d7688c307fdda66fd635c9bdbf4acd4707","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2641,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdpval_external.csv:row=10;column=Win Rate (%)","sourceMetricId":"Win Rate (%)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2641,"metadata":{"epoch_file":"gdpval_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdpval_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1075779a9b1a6008ce19","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-738a687ef467a99fcaaf04b2","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Win Rate (%)","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.348","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1411612e9a1ba31acf23461fac028ef4984efd44","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2638,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdpval_external.csv:row=7;column=Win Rate (%)","sourceMetricId":"Win Rate (%)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2638,"metadata":{"epoch_file":"gdpval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gdpval_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-99bf6fd065a1e64c87cb","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e814c902891ef5e80452d26","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Win Rate (%)","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.49700000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6230002dc12e94f16551188bf9c1e0e4e4a13aad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2633,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gdpval_external.csv:row=2;column=Win Rate (%)","sourceMetricId":"Win Rate (%)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2633,"metadata":{"epoch_file":"gdpval_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-geobench_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fa09d1a22780ccda65e9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://geobench.org/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e84ef536963e78b83526b6bb","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ACW Avg Score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_32K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2915","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83caff55433289a2c6668c5282e577ede69ad992","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2669,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"geobench_external.csv:row=27;column=ACW Avg Score","sourceMetricId":"ACW Avg Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2669,"metadata":{"epoch_file":"geobench_external.csv","notes":"Thinking token budget was 31,448. Rounded to 32k to reference existing model version.","organization":"Anthropic","release_date":"2025-05-22","source_link":"https://geobench.org/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2915.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-geobench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c2c64656e6594cfeb99f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fc64486fbcf4cd2e916a61a0","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ACW Avg Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"3877","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9bbcc3a6cc5620334a00a50ed88118b8029cd9b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2654,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"geobench_external.csv:row=12;column=ACW Avg Score","sourceMetricId":"ACW Avg Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2654,"metadata":{"epoch_file":"geobench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":3877.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-geobench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b19177888c2faeac1caa","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://geobench.org/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ad749005c6c2cc24ac83b17","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ACW Avg Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"3937","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cabda3bb070afce976ddccda74827be9fb4900df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2649,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"geobench_external.csv:row=7;column=ACW Avg Score","sourceMetricId":"ACW Avg Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2649,"metadata":{"epoch_file":"geobench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://geobench.org/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":3937.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca98dcdaec6ba77a8e4a","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gso-bench.github.io/leaderboard.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64f98ddf3b0e6cf2567ff52d","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.049","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6836617e75de2ca7a9040ed9baaa51430b5cd88","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3137,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=28;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3137,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":"https://gso-bench.github.io/leaderboard.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.049,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a5b3b52a9b8f50aac4a6","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gso-bench.github.io/leaderboard.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5095df32b7920a77cd241941","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.069","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"132fda852a5deec616784c46345e08a14a70a9af","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3132,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=23;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3132,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://gso-bench.github.io/leaderboard.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.069,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2227e9cf2246c030a020","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7085d52fb24367d4a4094a66","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.06860000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"132fda852a5deec616784c46345e08a14a70a9af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3134,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=25;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3134,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.06860000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-29d1bc281233513cc4d0","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gso-bench.github.io/leaderboard.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64f54d890dd628ffeb8c7c1a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.265","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bdf30b59800b301d6e596e270ce173d9efc767e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3121,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=12;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3121,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://gso-bench.github.io/leaderboard.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.265,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5147d9d8caa781610f31","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6322ee69b61180c3062a1ef","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2647","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bdf30b59800b301d6e596e270ce173d9efc767e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3122,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=13;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3122,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2647,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-028c6864f4dbc12ab718","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5703d23524201c086c76c6d5","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b934673ec5b53c69716071c931f9144ad3b3e42","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3118,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=9;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3118,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-abd4603dcded2c5a169b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-431481067d67f28c9ed7bc28","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.412","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b934673ec5b53c69716071c931f9144ad3b3e42","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3114,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=5;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3114,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.412,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1df9920dcfed354d8000","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ec0052282fe352bc9fbce0c","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b934673ec5b53c69716071c931f9144ad3b3e42","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3117,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=8;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3117,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7f076db48a6ce77288e3","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30977a414dc3b419bf7b4f64","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4412","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3356417323406a8db3cb77e97435c5d6b3cb5eac","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3113,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=4;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3113,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4412,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7338596d3f9b14ea0cc5","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4a349aebcd7a7ab5ce37aebd","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4412","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3356417323406a8db3cb77e97435c5d6b3cb5eac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3112,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=3;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3112,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4412,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8de29a51c44c79248ec2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8830684e2bab63dfba159a0a","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4706","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2604ae6060c3c6af781f1176d2da8f5529cb646","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3111,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=2;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3111,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4706,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2ae2d82d539a389a3a9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gso-bench.github.io/leaderboard.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-344acec0bcf169072f1ba4a7","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.147","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f2422e369db5f1718dac90211d28e2f5a6ca854","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3127,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=18;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3127,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://gso-bench.github.io/leaderboard.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.147,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fead7d869525ddb5cf8c","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-97eae7feac8fd257b820e1cd","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1471","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f2422e369db5f1718dac90211d28e2f5a6ca854","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3126,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=17;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3126,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1471,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1b3ca9172bb2832eee89","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe4c6ea0691f0b311593ba60","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.3725","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7ddb15f6927874f6ac3a33440d54ae44ba5eafad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3116,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=7;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3116,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.3725,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-256a76ec739724b3cc98","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gso-bench.github.io/leaderboard.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b45188cb124c80071bc36559","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0392","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"01eb5fab8ad6960bd113de0c1dea9e041750ecda","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3140,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=31;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3140,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://gso-bench.github.io/leaderboard.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0392,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec3cb253bc44ec45b952","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6b176bad48dbb4de2c20b11","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2255","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6d3b391fca601d0b646c62e243c8292bc51faf2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3124,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=15;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3124,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-376789ab53c41230dc71","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://gso-bench.github.io/leaderboard.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ca0ef35faaf4557804e3b5a3","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.06860000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6181c55c9eb9c7217ae155c4dfa714e20bbc6775","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3133,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=24;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3133,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://gso-bench.github.io/leaderboard.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.06860000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-396c05c554ae1e9d5a64","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ee75eae96b0dbb2546679fe","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.137","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72ab6c2e6ef20ae956be2a2cea0373b3f2ff36b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3129,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=20;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3129,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0bb888da59542f6e8906","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5f1e833895d996b6a1a3ea37","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.1373","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72ab6c2e6ef20ae956be2a2cea0373b3f2ff36b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3128,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=19;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3128,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.1373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-11936d8b6319dcbf6fa4","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-200748eb7353dcf3ade4ff26","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2745","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71c4e41d93dd6a06ad834b128f0a54e22355940c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3120,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=11;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3120,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2745,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e81a2eea03f558303566","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9729a15f8d3524da16bc2676","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.2549","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a47bee525c34ecd0299f581522fd71fa93dd0c7b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3123,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=14;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3123,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.2549,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fd4347f366405ce948b4","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c9c41e4d3e0e36aa94a90295","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.31370000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a47bee525c34ecd0299f581522fd71fa93dd0c7b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3119,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=10;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3119,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.31370000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-gso_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e1581b5cf2ed5ce57f98","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-da7245d7fe9c8ca6449fa4c0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score OPT@1","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.402","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"46e5530d38ba2eca81cf8e4756ad70e80c1c616b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3115,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gso_external.csv:row=6;column=Score OPT@1","sourceMetricId":"Score OPT@1","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3115,"metadata":{"epoch_file":"gso_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.402,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-lech_mazur_writing_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a122ddd43de3d5a6b60a","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://github.com/lechmazur/Writing","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98e34c60f6c725281b1f8c9f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6f9e6a76a4ff35e87ae80213d85bf8796a1dead","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3410,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"lech_mazur_writing_external.csv:row=46;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3410,"metadata":{"epoch_file":"lech_mazur_writing_external.csv","notes":"Score from the final V1 leaderboard (0-10 scale) of lechmazur/Writing, commit 80b7f17 (Aug 8, 2025). Frozen: the benchmark's later versions use a different scoring system that is not comparable to this scale.","organization":"DeepSeek","release_date":"2025-05-28","source_link":"https://github.com/lechmazur/Writing"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-lech_mazur_writing_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b98123f3298bd9bdc2a3","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://github.com/lechmazur/Writing","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6f8af0565b66df237eb9e378","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc79c8e822a02fd22bde8809436ad30752d916e8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3408,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"lech_mazur_writing_external.csv:row=44;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3408,"metadata":{"epoch_file":"lech_mazur_writing_external.csv","notes":"Score from the final V1 leaderboard (0-10 scale) of lechmazur/Writing, commit 80b7f17 (Aug 8, 2025). Frozen: the benchmark's later versions use a different scoring system that is not comparable to this scale.","organization":"Anthropic","release_date":"2025-05-22","source_link":"https://github.com/lechmazur/Writing"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-lech_mazur_writing_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-22d73ae562d96cbd3dc3","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://github.com/lechmazur/Writing","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e8be7808dd51ed3e3bc4682c","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc79c8e822a02fd22bde8809436ad30752d916e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3407,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"lech_mazur_writing_external.csv:row=43;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3407,"metadata":{"epoch_file":"lech_mazur_writing_external.csv","notes":"Score from the final V1 leaderboard (0-10 scale) of lechmazur/Writing, commit 80b7f17 (Aug 8, 2025). Frozen: the benchmark's later versions use a different scoring system that is not comparable to this scale.","organization":"Anthropic","release_date":"2025-05-22","source_link":"https://github.com/lechmazur/Writing"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.36,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-lech_mazur_writing_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e7fd73f919aa6efdefc3","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://github.com/lechmazur/Writing","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e30dde770f48a02064efdac3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db4c2f3737595286e7b55102092a0ac0cce101ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3400,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"lech_mazur_writing_external.csv:row=36;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3400,"metadata":{"epoch_file":"lech_mazur_writing_external.csv","notes":"Score from the final V1 leaderboard (0-10 scale) of lechmazur/Writing, commit 80b7f17 (Aug 8, 2025). Frozen: the benchmark's later versions use a different scoring system that is not comparable to this scale.","organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://github.com/lechmazur/Writing"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.38,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-lech_mazur_writing_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-22e029b1e4b4c2b6bbbc","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://github.com/lechmazur/Writing","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6e489fbd6e1ab133864c3cc","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c040d8c27dccc8f044bdac0cdfc541e56f04565b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3397,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"lech_mazur_writing_external.csv:row=33;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3397,"metadata":{"epoch_file":"lech_mazur_writing_external.csv","notes":"Score from the final V1 leaderboard (0-10 scale) of lechmazur/Writing, commit 80b7f17 (Aug 8, 2025). Frozen: the benchmark's later versions use a different scoring system that is not comparable to this scale.","organization":"OpenAI","release_date":"2025-08-07","source_link":"https://github.com/lechmazur/Writing"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-lech_mazur_writing_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee928a4202c45b2a93aa","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://github.com/lechmazur/Writing","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9547c487a1c20af9976da4a0","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"8.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4776ee307c70981b53c0404d798623c110692563","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3409,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"lech_mazur_writing_external.csv:row=45;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3409,"metadata":{"epoch_file":"lech_mazur_writing_external.csv","notes":"Score from the final V1 leaderboard (0-10 scale) of lechmazur/Writing, commit 80b7f17 (Aug 8, 2025). Frozen: the benchmark's later versions use a different scoring system that is not comparable to this scale.","organization":"OpenAI","release_date":"2025-08-07","source_link":"https://github.com/lechmazur/Writing"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-96bd1a17728f0407b7eb","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-31e87cf3be22ae4481d00a99","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9663897280966768","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c13d2b087a2231701ea3de4acf4c3525231eaf1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3480,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=13;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3480,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.63897280966768,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0bd1be22fe8176601f3c","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-714e676ef232612ceaa029e4","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8691465256797583","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dfbd4d2a0b638eb29bd7c9fcf1221f11d81d9f8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3473,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3473,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.91465256797582,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc1235fcfb1a6c94dde6","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1cb2d090bee9b5b0dd9e1e0c","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9636087945413192","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dfbd4d2a0b638eb29bd7c9fcf1221f11d81d9f8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3471,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3471,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.36087945413192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e516448d049874e2f048","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-381f081247556ee0b7ed4b09","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.850453172205438","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6d01b20fb8a1c8336050ffe492d092e114929a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3482,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3482,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.04531722054381,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7963dbf2e21b6e4d5e55","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a6d384de420f1ae339e18343","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9773413897280967","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23534e62f8814b15d6adc0c9277cf2b684124bc9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3472,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=5;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3472,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.73413897280967,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d07657bf0a24c3256c9f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b3486502e42c67b6c9708e19","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9813066465256798","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2c39ef917ad22b6e0e0be2d9617d94d165f738d3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3470,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3470,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":98.13066465256797,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aaa9e408363792a58d10","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c450de1365772e2fed9da768","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9792296072507553","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2c39ef917ad22b6e0e0be2d9617d94d165f738d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3475,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3475,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.92296072507553,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2526afba9f0026792226","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed5d0e6d09c77682462a5882","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9784743202416919","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc0c5f0b8891239640f8502991e4deb19e03ba5e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3469,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3469,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.8474320241692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d6493d0b63f7f7da8eb4","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b9b2f0331b3dd8fe465435f4","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9679003021148036","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc0c5f0b8891239640f8502991e4deb19e03ba5e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3478,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3478,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.79003021148036,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6eecf63e41431057b5ab","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8c9c06b3f0e806c925cf7252","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9490181268882175","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb44912cbaed7c9ce2e8e7626a3560d2ca666d6c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3477,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3477,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.90181268882175,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-math_level_5","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3b6cb8152a1d127fcfdc","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c23e79fd1349eb586df8b47b","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.952416918429003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb44912cbaed7c9ce2e8e7626a3560d2ca666d6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3476,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"math_level_5.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3476,"metadata":{"epoch_file":"math_level_5.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.2416918429003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-97e6a7d2193565cd7f2e","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7630393b6ec5995a7cb89f4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.537826","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e900b40479f805d5c8c3e98580234fa48fe7e92","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3606,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=31;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3606,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.537826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8155deb7c6562dc8298b","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-02302de729d2b6b717b24d9a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.639292","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7252e05078f4a0de536c55fcc2737e77e9bbd84f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3597,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=22;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3597,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.639292,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9c1dca2c8f85b1d8e8ec","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-42fbe2db4fe495fd38df20d5","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.615452","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7252e05078f4a0de536c55fcc2737e77e9bbd84f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3595,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=20;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3595,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.615452,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-847a638218eb7f9ea86a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d26614c69e8e5949634a028","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.73007","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"753c078c13c834a74f264a476e495445a46f351f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3584,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=9;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3584,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.73007,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8227da55d12fadcfd9c4","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7d9b45e0427b8dab9ecb4c9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.74969","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"753c078c13c834a74f264a476e495445a46f351f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3585,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=10;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3585,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.74969,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f6651000603a777018f0","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c8e5c398ec66bef427cff81","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.788644","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62d5030f39664b53d5d9aa8f65f242922f9c716a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3578,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=3;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3578,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.788644,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f01c37caa805601261bb","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b01cfba365add511e723a1ae","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.673842","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ae10b166c8d9feb950f1df4f50aedc9aa5da1f7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3590,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=15;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3590,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.673842,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d07b82bd9210071b9fa7","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3fd099beb13d63e1c2a2557c","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.77035","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2acbc201aa8e1d2a872359a1ebcc6d14ed731beb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3579,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=4;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3579,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.77035,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8e1ed47b69b4f57c48d6","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f15eeb8a79cae7a32d282357","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.693697","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ae92ebc0bbfc8cc0088f1b98dc4f1e38e37968f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3588,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=13;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3588,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.693697,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-de5a776cadfa64644ec7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a168a5476f62cfa143e7b55","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.696053","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ae92ebc0bbfc8cc0088f1b98dc4f1e38e37968f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3589,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=14;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3589,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.696053,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-11-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f8cf13b7344190e72d91","canonicalModelId":"openai/gpt-5.1-codex-max@2025-11-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-26f92cf67e0c3475d3ca4af1","mappingCandidates":["openai/gpt-5.1-codex-max@2025-11-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Max release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Max release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5.1 Codex Max","modelRef":"gpt-5.1-codex-max","observedAt":"2025-11-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.708301","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3c66999bfb4d18d7f7be422362c24cb697e51e9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3587,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=12;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3587,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-19","source_link":"https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.708301,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4bffec8abf1ee56aab41","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ff58d445dd61b2eb6401e49","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.752875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3afa9b0f88caed0374b4ead94c328ef19b6115b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3580,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=5;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3580,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.752875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6c8b1e9dde2f3758e165","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4d1373a61ca0a5d96e1658ad","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.745439","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c514d5d6d0e6aae2f47de16825e9d8b762a53042","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3581,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=6;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3581,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.745439,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-65af66ea6791d27be169","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f68c33cee2761d974f02a9a9","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.743443","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"379bcb91f7e023094b3b85ad9da7ae773fc310ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3582,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=7;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3582,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.743443,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-metr_time_horizons_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a65ded66ebb953d78651","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-49112c8e4356b027a0e9811d","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"average_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.743443","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"379bcb91f7e023094b3b85ad9da7ae773fc310ea","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3583,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"metr_time_horizons_external.csv:row=8;column=average_score","sourceMetricId":"average_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3583,"metadata":{"epoch_file":"metr_time_horizons_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.743443,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mirrorcode","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a2eff3a77874bbd8b6a6","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f1cde3caef874d8a450c855b","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.638888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21719fe8460431f23dd8cda476f040bc8de8130b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3637,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mirrorcode.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3637,"metadata":{"epoch_file":"mirrorcode.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":63.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mirrorcode","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d3385c12ee8f2cac04b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a015166c17a7d8eadb61e814","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.3111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04b29f7f6f625d0e2bc8e7289dc7887748d532cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3632,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mirrorcode.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3632,"metadata":{"epoch_file":"mirrorcode.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mirrorcode","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-90ff10daa934615566f4","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac1f261f9f047aa68789c2e7","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.08888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"593461745065496aa7e5647cb68365cc7067d1cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3633,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mirrorcode.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3633,"metadata":{"epoch_file":"mirrorcode.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.888888888888888,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mirrorcode","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-923c2d757b9807a804b1","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3d2c84e5299a0beb9b186e4d","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e75cc85b2568b43c9283b9759093a37653320c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3635,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mirrorcode.csv:row=5;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3635,"metadata":{"epoch_file":"mirrorcode.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.555555555555555,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mirrorcode","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4221af1c138950e4c883","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a170508a39f0c79e6a0553fd","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ab8aeefd7d061eecb3d7e1443713e717a5240b9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3634,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mirrorcode.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3634,"metadata":{"epoch_file":"mirrorcode.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mirrorcode","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f371a39e5bb39163ceb7","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9d8a8406a16d629cb107db08","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.2","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"769be88d36168cf55d67cd306981040201a1d5f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3636,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mirrorcode.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3636,"metadata":{"epoch_file":"mirrorcode.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-de1847532f68dbc96fed","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b8871333dc7d4a7295b835ed","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4bf6fda326d6274d48ec7db5dc2074462ab4aa7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3888,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3888,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":52.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d17ab44ac406dff0b206","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f828e21606f96aff62b2a76","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Epoch context-window suffix on the dated Claude Opus 4.5 release.","registry":"model_aliases.json"},"mappingNote":"Epoch context-window suffix on the dated Claude Opus 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_48K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dcd4b59d77b34844dcd19cf9c3725f6d50a0732d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3901,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=48;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3901,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e689006a2b3d1d802e0f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ede51262c9797192b0e67c91","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"661f894193126b6c8f0e7b80c98093639676f264","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3864,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3864,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-88f3464af0f385d987fa","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0c84c680979551c815ef803b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_low","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"661f894193126b6c8f0e7b80c98093639676f264","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3865,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=12;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3865,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f12a3659f9c7bed7a7fd","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-896043a25b526b0b8ce8aa88","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"661f894193126b6c8f0e7b80c98093639676f264","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3904,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3904,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e2c1916d36d83a1b9a2c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f8d5018ec2fbe8724c71e376","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"857770abed152604d416fc7f185404ff1380d04d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3863,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3863,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1d2701598803d80748f1","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c34073b8c1b7fcb8720ab90a","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.28","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"857770abed152604d416fc7f185404ff1380d04d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3905,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=52;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3905,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-501581fe8c8048242a9d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-48505c858cce5b488f432e6a","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4f34b33f4b5023ace990b293959cdc5cc08072d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3902,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=49;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3902,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1ff11aef23798644c5e5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-351dcb8297fe8c4d0f69732d","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4f34b33f4b5023ace990b293959cdc5cc08072d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3898,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=45;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3898,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-090df4a9604b4b1c4a2a","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d24321639de37ab4c4557d15","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"014e1d191a45ef7ca32ed6c77ba761c83fc485ff","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3862,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3862,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-063cd45d35b53f59b581","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e520c159ca0ab76a888a61c2","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"014e1d191a45ef7ca32ed6c77ba761c83fc485ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3903,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=50;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3903,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-51668768a3c90567078a","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a0814b6324eef39889db2585","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_48K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da0e1bf200ab49d320936978a7a71dcfe3c749f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3899,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=46;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3899,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4572a99a3f550fdf1a78","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3599a33d7a37a408e3d3fa6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeb5d4b6e0b60ad41128a909545a163530aa1655","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3860,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3860,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.000000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4c86a3f1e874ab4b37ed","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-326d1c13001d152a8d8ef59b","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_low","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeb5d4b6e0b60ad41128a909545a163530aa1655","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3861,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3861,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-93599ba7701948635f94","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0c9ee482ea6bffd2a5afd405","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbe95bc01f8457c675b53dad409ba3d17d76e057","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3869,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=16;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3869,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6629fc35af12aa48c8be","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e09fbdea4824f9dcd784bded","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbe95bc01f8457c675b53dad409ba3d17d76e057","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3890,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=37;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3890,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5157452d3be8332161f3","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3bd9e637e80089275248478c","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.34","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d596e4c675926f4e08b159c190db9ee5afb73b11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3879,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=26;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3879,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7f6989c5d8d0dcf08220","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8e2ebd31719c47011aa49dbe","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d66d9eaabd4917e67a796def3750b81e4005f315","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3855,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3855,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8bf633710c314794598c","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ea3df8f1b2087e83f030a800","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.34","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1979dca314aa2b706a34ef4fc91eb53932754aec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3896,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=43;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3896,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b35a6fd6b54cf80e365","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1e2458022262a9398b21ea1b","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_low","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1979dca314aa2b706a34ef4fc91eb53932754aec","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3868,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3868,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.999999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7f9dd589a13b16b48807","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7dca6686ab3458c80dfcbbfe","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_medium","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.32","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1979dca314aa2b706a34ef4fc91eb53932754aec","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3867,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=14;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3867,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a92cb556150fc626cd92","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3eb3b2f8563033aa80590476","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c15fcd0c497f5d22ae4d292357311644286854ea","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3878,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3878,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-462ff3f47e5d21261af9","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b49a909f2cfa7e61eee9f344","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c15fcd0c497f5d22ae4d292357311644286854ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3877,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=24;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3877,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-73b3a676f774b5330b5a","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6c15df14f27d0e2a92468d70","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.32","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bfb322e3cb20711060d3bb07068f46fd70147874","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3897,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=44;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3897,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-35b77471c72448f4af3c","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd6ca5e6f952ece3621361e1","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_low","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.28","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bfb322e3cb20711060d3bb07068f46fd70147874","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3871,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3871,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.000000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3a9a7fc634b26390a301","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-90714972ed52eef7eca6dfc2","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7733c453c10cc6135e1ab9c4e5e4d062453ae1f0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3886,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3886,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":30.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d607c69c685fff50f286","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b3923c9fca29c0ee9d78ab40","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7733c453c10cc6135e1ab9c4e5e4d062453ae1f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3873,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3873,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d6bed68cd96d1231e06","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4ec88da5986f25abf8570104","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7733c453c10cc6135e1ab9c4e5e4d062453ae1f0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3872,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3872,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5fdee7ffe4b24c3b9326","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bac7e4505c81430925ddb48c","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f12c83e1b5bb044908a948d872d5a3e1ee30cc2e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3857,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3857,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d4ffd0b067a4b2ab577","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dcef9d33dcfa98e367b23b70","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6a69c0981f282d991021f6e1ca490369912244f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3885,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3885,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7a4b491928ef3fc7b747","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-af807cba791b4a93350164ff","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec9de81a9ca6e9c26365269a3bcad92a980f1657","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3859,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3859,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d53c6574e1f4cd328262","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fbb73d158f32c5001c24494f","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.37","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e034b9a78109d389099c86ddc67745339cd64e7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3906,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=53;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3906,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d32b6e15bce921303543","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e33ecb83284eb9559a069d0e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c696c546ad5c4b8aea8c2ac2f0974b460ba021f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3895,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3895,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":52.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1683d7eee552c7e553ed","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6b9037508832c46568dd0b4c","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.32","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c696c546ad5c4b8aea8c2ac2f0974b460ba021f5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3866,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=13;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3866,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d92238de79c5ac6d7595","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ce140fb69199623be0ddf816","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.56","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c696c546ad5c4b8aea8c2ac2f0974b460ba021f5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3907,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=54;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3907,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.00000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-09a1f81723e4965f5cc1","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b7eecfacfafcf204d84270d8","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efc376d17cdcdccf4ac1f8a48994533e546d2801","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3891,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=38;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3891,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a2094a16e00dc03a2c4d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a5403f685a01369cbb507272","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a5e3b62d683c8ab8eaa6f739dc5e329376587f51","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3893,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=40;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3893,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e1157eb4fc918845cce6","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-537c35b8f6eaca0ba92a350b","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af7cbd0d2a32a519c04dd4e6676d86c509f1e308","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3892,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=39;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3892,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a8e39353bf7e7c499b78","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dfbe040e14ec9829b898af1e","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.34","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb3f840046064dccfc712eb0b4fbfa1049ceb6a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3858,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=5;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3858,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-880631a1c1478339548f","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c95a8e44cb7d3da0d28bf7b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ff30f42f4f7966b5dbe710a48c8511096d7f77ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3908,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=55;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3908,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4b3f65910f5e51e2ca18","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8efe0299b70261590cfb93f2","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df6fc0b8d25aae687b71686d41a223fae76f0ebf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3889,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3889,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a24718be52d55b2606b1","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac3c8f604eea00260ce7f48f","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash_none","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3619a73528c3e8fc4d2c78d8e5f17c3c6344b7ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3884,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3884,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-635a74d3e289d63d0d7c","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4cad57f18542e9d96cbfbca2","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57cc6804d529f07606b6e2d87299fcfc564c78c1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3882,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=29;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3882,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-958d3c25dfc68a87d503","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77382043352da4e045d8f9a0","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus_none","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57cc6804d529f07606b6e2d87299fcfc564c78c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3881,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3881,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dc745e7691e68da75d8e","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-67ceaf93c2ea74bed7bb53dd","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash_none","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"096ea66f7fa3e59f2a959fb8a104cb27541b56ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3883,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3883,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.000000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-da907746ec8469dbff2e","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-164503134b5941db3ea6337e","mappingCandidates":["qwen/qwen3.6-max@preview"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on the explicitly preview Qwen3.6 Max release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on the explicitly preview Qwen3.6 Max release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview_none","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03e11dd356cbce1113f5fc3176e6b543b0baab88","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3870,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3870,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-683cd0a0b1aa35e41e9a","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6625fd53a7ab7d53b3041f6d","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus_none","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.27","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"325a4db20b8effac69c27dde71d7411b07614388","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3880,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=27;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3880,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-605a18ca62e258458ac0","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-106b9740e89c59e358b46de1","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.32","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"16bb981f434712a2a716c8053f9e5667c3ed66c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3894,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3894,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-mystery_game_puzzles","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2879f4d69a72fe89024e","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4f0497bde3740ca9b4b113a5","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e55e5a5bf11b7b8713827d8128a2341fb44c793d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3887,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"mystery_game_puzzles.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3887,"metadata":{"epoch_file":"mystery_game_puzzles.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0fa22b363a78e0bae832","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a7f0f685eb60f5059f5c6d48","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.046","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"98c35929712a27c4c52e5449e0e5b926a2cd08ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4006,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=9;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4006,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-751962a196af51f9d80a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9df4b62d2ceee7eeccaae767","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.182","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7db8b0bb1571d8c62899b1d649a80d1ea52b2e46","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4001,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=4;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4001,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e57d1152d938821e3bb","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9091f9419adb3fd6e655f55d","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.139","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7db8b0bb1571d8c62899b1d649a80d1ea52b2e46","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4002,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=5;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4002,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-507f84e308fa46c6f16a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd9f7c4dfcdfa7b902ce664e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.1852","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cac719117bdd4603a1a910f784c55fcc14fc66ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4000,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=3;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4000,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-daaab01c0bf5ad5479d8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-acc178e1c0386dd7425af869","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.20600000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cac719117bdd4603a1a910f784c55fcc14fc66ce","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3999,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=2;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3999,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cd20839a8684c8e5488a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a3a4bcc00a53603e20b20d6b","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.083","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dca0f40388c93f657f9eff8fec19b84b98a3c015","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4004,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=7;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4004,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-14943fdacdca9799aa8d","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b3adc8bbed0674f31868a091","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.09300000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dca0f40388c93f657f9eff8fec19b84b98a3c015","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4003,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=6;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4003,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f06b1a2db2b918ab39ea","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9c0bb7ad43b7dec7fdaeac72","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37a5ace261fc49bc300b774fde28c84f4e9c382b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4005,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=8;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4005,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b5ecf891ed34db66f536","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a49b3286d650378a4540b7de","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.046","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2200f62097dafac0a7fb8243ba2d5efc9fc016b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4007,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=10;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4007,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-osworld_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e717e8fbd32ecb56440c","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8daac7992c796157877bceb4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Binary accuracy","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.027999999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"290a1db128b5030d95f7801bdf78b5f896f25550","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4008,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"osworld_2_external.csv:row=11;column=Binary accuracy","sourceMetricId":"Binary accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4008,"metadata":{"epoch_file":"osworld_2_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9fbc50e6ca03e7c2dee6","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cbe8150ac2226bd252a40115","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6638888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3305c1c682f9a16a6446edb41321ab4b623e8d1a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4168,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=161;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4168,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.38888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2979694793566a5188fe","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-904bc75de143ad60661c1376","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5266bb0a09530b8a44dddad9b3340558744de569","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4015,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4015,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f69a203eb461b7826e9b","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-93c45c10c3a808f729ab87c5","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"1.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3307317b4e21140f1034da58be96cbf84d1bc1ce","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4079,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=72;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4079,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cc078806dad4409e6cd5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d2d7d20f6b1495b0e41994c2","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3307317b4e21140f1034da58be96cbf84d1bc1ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4078,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=71;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4078,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-65b25c671a060d6b2a58","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-920a584740d5001f4a34984b","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9972222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3307317b4e21140f1034da58be96cbf84d1bc1ce","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4114,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=107;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4114,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":99.72222222222223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-563dba2dc1791a24843d","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3e7b3a79713860f819342b9c","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35833333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"237e2cdcfda68d88947c998afc5d8aeb94e8fadb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4158,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=151;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4158,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.833333333333336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-78db8fc475d8858038c5","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7edaa1609344e61db02ac356","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6666666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"237e2cdcfda68d88947c998afc5d8aeb94e8fadb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4156,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=149;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4156,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.66666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8de9d296c02038a7bc2f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-05d1473674b8c4ccb5518803","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4222222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a67938a67023ef139fa9aca7fd0cb1cb3ee8f96","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4174,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=167;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4174,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.22222222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fbbe2a8a036e80a0c07b","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d19c7c3a2dd1d1414d283395","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a67938a67023ef139fa9aca7fd0cb1cb3ee8f96","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4172,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=165;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4172,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3996618b54c909d0542d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c1457b2a5308f0b087a641be","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_27K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6444444444444445","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a67938a67023ef139fa9aca7fd0cb1cb3ee8f96","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4169,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=162;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4169,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":64.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-43b528d40575aa8b0fda","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-60f238687cdb1a25c471409d","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.48055555555555557","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f89cf10cca39aa7026bc21834453b3bc2276a02","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4141,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=134;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4141,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.05555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fb5a93c64081c0e7cd73","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ae050b154b7655b1d02a7737","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8166666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f89cf10cca39aa7026bc21834453b3bc2276a02","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4143,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=136;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4143,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":81.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eab442b123c0477b3d54","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7dc6afa1be23e6e11bf3d5ee","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8611111111111112","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f89cf10cca39aa7026bc21834453b3bc2276a02","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4142,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=135;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4142,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-76d8e1caad0dbdb20cb8","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f3f27b561d8b5179a45f0625","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9305555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bb24e951800323be48150497026ea1a54755423","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4129,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=122;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4129,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.05555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2fe2d53496cad8a39b97","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a1b1a4726c0b113c58fb3d67","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_64K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bb24e951800323be48150497026ea1a54755423","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4128,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=121;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4128,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-281eb0d9fb3155b65942","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80d5523c6f734f505539d352","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bb24e951800323be48150497026ea1a54755423","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4061,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=54;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4061,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5df1d42452c05343a23a","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e454cf618347a5ea05524550","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7231af5874ae0a03d501444b7d10aae20dd80aac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4077,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=70;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4077,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2e3d9c39deb38a6f4672","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9c91aeb4dce6340d223b41cf","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.978","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7231af5874ae0a03d501444b7d10aae20dd80aac","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4120,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=113;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4120,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-701fc1fa346e3b68fef4","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-daf568826a7f6765be800cc9","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"39db32ff6d48aed1d0eb2ca58df31a501095863f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4076,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=69;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4076,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-74fab56feedf9435df82","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fb632add60450720797bd5a0","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9833333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"39db32ff6d48aed1d0eb2ca58df31a501095863f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4115,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=108;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4115,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":98.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-688453beaf5d98df2316","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88cc73e547df77f32c0e8f89","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"39db32ff6d48aed1d0eb2ca58df31a501095863f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4075,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=68;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4075,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc2e2bb18a8b0ee0c26f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5e5389480102410ed0dcb145","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac9e72d5b2503051dfbf349b5f0ecae692f4ed46","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4074,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=67;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4074,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3dcf9dc6f404a277ede1","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2b746311d780525a0b362f9d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac9e72d5b2503051dfbf349b5f0ecae692f4ed46","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4072,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=65;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4072,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f2cf84d5e8da4ec794e","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eb92fa3faf5039bc0497e02d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac9e72d5b2503051dfbf349b5f0ecae692f4ed46","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4086,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=79;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4086,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":98.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aaa2f08146e3001dd464","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e5a2bb57c78582745eb616a","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35555555555555557","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eda8ea90091172e0c0fd442edd2aee93c05661f5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4160,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=153;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4160,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b235287c9bdd4ac61f0e","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cc65443da9c10315cf43e232","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eda8ea90091172e0c0fd442edd2aee93c05661f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4153,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=146;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4153,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ae0f4a2db044a56bf1f6","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98c71544f8df6299c38d0092","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7777777777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eda8ea90091172e0c0fd442edd2aee93c05661f5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4157,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=150;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4157,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.77777777777779,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-765b31767906d86bbacc","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8cb0f9904d39ffedc08cf48b","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_59K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7777777777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eda8ea90091172e0c0fd442edd2aee93c05661f5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4152,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=145;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4152,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.77777777777779,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-475492d349a4fb51a099","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bf022e6d1f3873c7c4dd9bc5","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_32K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.857936507936508","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"279688fc63d92ba6c2cbce32131616cc309e4898","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4125,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=118;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4125,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.7936507936508,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-421d28e5f5da99358d5a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ea52e4f600a048b96027aa6c","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7555555555555555","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"279688fc63d92ba6c2cbce32131616cc309e4898","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4100,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=93;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4100,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":75.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-98f1cfddf3000132bbbb","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d37304988e08fdf0b07b2cfe","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"279688fc63d92ba6c2cbce32131616cc309e4898","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4073,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=66;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4073,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bd0d79bf38686b786f4c","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-73be04a8b7af18adde0020cf","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8222222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"279688fc63d92ba6c2cbce32131616cc309e4898","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4103,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=96;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4103,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.22222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e87416ff99682d54868","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ca24ba78de92cdf6e4b4cdf7","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d4fd1a92ddace9bd2b3cb508f245a7aa5470e90f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4071,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=64;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4071,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ab5ccbf1a04d3a215848","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-175bc348874a5d0ef050d8eb","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9472222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d4fd1a92ddace9bd2b3cb508f245a7aa5470e90f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4110,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=103;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4110,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.72222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6ef6f6256341754074fe","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dc5c043f359719726646f92f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V3.2","modelRef":"deepseek-chat","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bb51e1e68c60878ecaa6b51ee492792ded650e16","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4090,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=83;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4090,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.888888888888886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca593c923d4f506ed71c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c547176c94bc5a0de9075701","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f4ac10ce05897bec904f2a797ae25dd3ac9e1f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4084,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=77;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4084,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1f06c4a37064026a7185","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9634ca508e76139e97ccd735","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9861111111111112","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1be51a0b00fa741ed85486ce543892432dc26e8c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4010,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4010,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":98.61111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e40bffee1615ded2074b","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bac08a00cd22779cd1ef7b7c","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc3f4555d5185837b1d5d00e66d2cd04a8d5a614","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4060,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=53;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4060,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-285a825f02dbb396ff3b","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-888e3200392d553ed882bdf9","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc3f4555d5185837b1d5d00e66d2cd04a8d5a614","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4113,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=106;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4113,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f1737bfd753dea26e3ec","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-63c1c73d826938ce1d891ca2","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc3f4555d5185837b1d5d00e66d2cd04a8d5a614","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4059,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=52;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4059,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.666666666666664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fad504626667fbe5423d","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1a82c1a950899d1f91643742","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.922","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3fe1fec1dd90072cf7643796cb7c27d743b72222","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4130,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=123;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4130,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":92.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6de3e9beb821a602e9cc","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a98447da8b1bf20905ead5c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8416666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78fae8469877ec7c2f118c4ff2992ef48db2a53f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4146,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=139;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4146,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.16666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9cbf23307707a9611ccd","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-067ec002cadb837fdba17c31","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.956","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dcdd3ce4abe9ee8809d33bac40c10b51f5a3c9bd","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4126,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=119;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4126,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-97b63913f03a241758f0","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0fe00ffe7de86505206c55b8","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dcdd3ce4abe9ee8809d33bac40c10b51f5a3c9bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4058,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4058,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4849df892bb73e06b23b","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6867a313c64a644b5b7e63b5","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28ebeeee055973becb457a0ba1c377a3e4b39bf4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4065,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=58;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4065,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a218286f02297ac9197e","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-39aeba4be7bdee7371ac9c66","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28ebeeee055973becb457a0ba1c377a3e4b39bf4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4063,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=56;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4063,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04719450ff8d11a50b41","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-094690c94be27826d252fc21","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28ebeeee055973becb457a0ba1c377a3e4b39bf4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4064,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=57;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4064,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":51.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-85de7ab8aa8516fee5cc","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa03db29fb19e0819bad75de","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c817ed81ce21f94798022ae089622a49cd8a8e3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4116,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=109;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4116,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-81c7b4d71f4600c027a7","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-538314937c2733496441efad","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_low","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c817ed81ce21f94798022ae089622a49cd8a8e3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4066,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=59;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4066,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-134201d2b93c963384a7","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32ee25b28bf7363bcac8ebde","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_minimal","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c817ed81ce21f94798022ae089622a49cd8a8e3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4098,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=91;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4098,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4d7c30f7e0a834e78b17","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-acad1a7307fd5da6c373e34f","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9416666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"188bc30d944340287a64bbd846b353fc2bf37a98","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4085,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=78;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4085,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.16666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ab9ab02680e8305f6fbb","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9eedbcd873e5772d174140e5","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8222222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"188bc30d944340287a64bbd846b353fc2bf37a98","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4043,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4043,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.22222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6adf8afabda52d56dc44","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-27619cdcdd406ce3b83ed3be","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"188bc30d944340287a64bbd846b353fc2bf37a98","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4042,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4042,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-015597221ba49c90ee51","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2609b588181aa78b453b7a83","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9722222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8220a26f4e100bf0de8be7ecfd60d21f3a82e6af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4011,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4011,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.22222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-480f851ac17f0c9c6072","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-27c3401a344bb4ae8f3b170e","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8126a08f01f92f2a588a6f64c8fc74035676ee1a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4131,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=124;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4131,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3591e1158a16ba2b4ac2","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8da2c7d979b433f5ea986167","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cc578c5319df12556579ee5a4423e0ff37d40f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4127,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=120;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4127,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-09fd8830a90c17eff3f2","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2316afa4ae5f4d0bd7fe8f72","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f81388b5f65ae9fc5a9f48d8e3a5206615984a8e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4017,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4017,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4983643cd37646746244","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64529557d3137cd182518bb7","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8638888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"38337984245bcce9220a0df0ccad84db8f205fa9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4111,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=104;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4111,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.38888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-66ce4baa3faea7c368d9","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f934db767eeddfe7ae88d228","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.28888888888888886","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"38337984245bcce9220a0df0ccad84db8f205fa9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4016,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4016,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.888888888888886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dc02d5eceb08a4d23f0b","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-905694388eb16ed126f05d6b","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.3","modelRef":"glm-5.3_max","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"604199a0a9f08f1927d1bd214376891a5478d2e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4009,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4009,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8e8a44975c6f2bb3712c","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2fcb5614d7e99e2b82d9fb3e","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9138888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a6fe69489cc51b8e7b56acbf10089d0b137959f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4151,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=144;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4151,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.38888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0a6c59ded0b2e09a1d19","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-de8b835dc9482adf66e5bcd8","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8722222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a6fe69489cc51b8e7b56acbf10089d0b137959f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4161,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=154;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4161,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.22222222222223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-91011d0c9021e4a7a884","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53a8fdfff18da7b7a154d2e8","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a6fe69489cc51b8e7b56acbf10089d0b137959f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4088,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=81;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4088,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.666666666666664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3ea9db6bcb09f72e2295","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a43e02d5fb4e1725ac9b7c4c","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a8b6d764a903a5d137ccf07a7e39cdf878d1815","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4150,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=143;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4150,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2a80275c69f26da1aa25","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b8dc75cd7b4e06fc1875dea2","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7833333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a8b6d764a903a5d137ccf07a7e39cdf878d1815","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4162,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=155;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4162,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-905cb87a2c271a1463c5","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d01c5826cc808dbd93c84d13","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a8b6d764a903a5d137ccf07a7e39cdf878d1815","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4055,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=48;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4055,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-570ab4f1881107c6e6de","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0c21237eea6749c9707ed375","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ae9cdf61f0b42d65f525d0dfd716caa1a39dfa8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4149,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=142;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4149,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":81.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3ab0ba35fb74343620d4","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-436c026dea6e34226ffdba17","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ae9cdf61f0b42d65f525d0dfd716caa1a39dfa8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4099,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=92;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4099,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.666666666666664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-52281872ef273734fe6c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-963eb5d76022e43542fa4bf5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7416666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ae9cdf61f0b42d65f525d0dfd716caa1a39dfa8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4163,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=156;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4163,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.16666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-647f277e192ed6414b77","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b679b26388401b5e5b96f014","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35555555555555557","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ae9cdf61f0b42d65f525d0dfd716caa1a39dfa8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4101,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=94;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4101,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b2d78b11eb210a8e313","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8013d5e8c7cc0cb2f69ff5a5","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8861111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6cf7845d36b5d54dbd1a7248f683cca3b7324ef","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4147,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=140;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4147,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.61111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ded923c950a9d34aafe4","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d42547490f97eeca5403afa","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6388888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6cf7845d36b5d54dbd1a7248f683cca3b7324ef","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4140,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=133;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4140,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":63.888888888888886,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c3c41d6a1ba85bb4893b","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6d16b9e45558d610b250b7e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8555555555555555","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6cf7845d36b5d54dbd1a7248f683cca3b7324ef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4145,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=138;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4145,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2e7c82435b2b347a97f","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-94efb4f3d917f1822cf59d88","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.37777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6cf7845d36b5d54dbd1a7248f683cca3b7324ef","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4030,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=23;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4030,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.77777777777778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f21c7a8e9319988dd3ad","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f58fb5201078b8eeff809b1c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9611111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77f3a57769af848def48961d10c7388018e43f5e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4137,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=130;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4137,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6b1e2b54da96b1bfe414","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-11e404fd5ba034db69496290","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77f3a57769af848def48961d10c7388018e43f5e","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4138,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=131;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4138,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-112a406e517d9a8871cf","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3d5978a2c59e0fabde151c9","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9388888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77f3a57769af848def48961d10c7388018e43f5e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4136,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=129;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4136,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ef6a3cd4d8e274e7f0e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-693716815216496d64a2bd0e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6222222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77f3a57769af848def48961d10c7388018e43f5e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4102,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=95;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4102,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":62.22222222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f1441573ecc25d3a568","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7e19c18b0f768ff458e408a5","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.961","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77f3a57769af848def48961d10c7388018e43f5e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4134,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=127;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4134,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32fd4ce2cc27cc2f9300","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec87d47b6033e2f40c65f27f","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0728604fdb93e2221e4fca6ef7432e983355eac5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4096,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=89;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4096,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dbb4ef03566c117844fd","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-edf993eef5d2da52be012b83","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0728604fdb93e2221e4fca6ef7432e983355eac5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4097,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=90;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4097,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-40e1b725d000f39259b9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7202fe4cce7d9a5cb07c59c2","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0728604fdb93e2221e4fca6ef7432e983355eac5","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4094,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=87;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4094,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ff872b5f23557815593","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8db681b650d7d3d914555127","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0728604fdb93e2221e4fca6ef7432e983355eac5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4095,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=88;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4095,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-190d1b23264af45dec9a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2b356bbdc758fed09fce5f5c","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.953","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0728604fdb93e2221e4fca6ef7432e983355eac5","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4124,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=117;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4124,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b461e148f18261db0997","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9907c8a71eed8446c62a7975","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8722222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a379e7917affe620f49d29fa8440dd052d716528","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4121,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=114;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4121,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.22222222222223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bcff78627604069102fb","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4cd8ce94901bf53274cd94ff","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.26666666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a379e7917affe620f49d29fa8440dd052d716528","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4024,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4024,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.666666666666668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-52157688469a9ee7b8d6","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0031bab9d18ff42a11def72a","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a379e7917affe620f49d29fa8440dd052d716528","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4018,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4018,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3a53f3762f356c2a70d8","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-43accd713f83ff77a5ca3600","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19055f094b1f947635aef6a275363a7ecf105924","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4122,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=115;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4122,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.78,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a9e53febb1d360f40030","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c625a5137a0999271e095dc9","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19055f094b1f947635aef6a275363a7ecf105924","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4056,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=49;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4056,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-de6350b4e386e839f712","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba483ffa69cd3ee79f874dbf","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19055f094b1f947635aef6a275363a7ecf105924","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4052,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=45;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4052,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.666666666666664,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2b6c9b4cd3858a80c087","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0c720a723096512907a6f90","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6805555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2577948440aeebfe6f007adc7d2d02a412535296","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4083,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=76;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4083,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.05555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c736b8d28f22680247be","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3efb2755e63833ca444e4ad3","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2577948440aeebfe6f007adc7d2d02a412535296","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4027,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4027,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec8397c2aede6fab0b37","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-81b3ce08a7c9db0a0498ea9c","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2577948440aeebfe6f007adc7d2d02a412535296","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4020,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=13;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4020,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fe32daba8608f624fe39","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-35a9b6f94936dc7be4556d23","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6666666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"324ec36f7b2ef2acde4136d4f8607b417f140070","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4050,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=43;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4050,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.66666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c5bd8bfcd1724c16de6c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-267e796f9cc784b8e1d9b562","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9833333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"324ec36f7b2ef2acde4136d4f8607b417f140070","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4107,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=100;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4107,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":98.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2b574b8adce9460bc768","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a6771c14c16992760ae2efce","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"324ec36f7b2ef2acde4136d4f8607b417f140070","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4034,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=27;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4034,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-10d4c38e71be5e893719","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8be56ab7d484fcd3121bfeeb","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae80ffefe9422f8f5516ac55971b6def5fe58ab1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4026,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4026,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6d5728077e66f102105c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8874657f7e6840d5f9d8b0a9","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"1.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae80ffefe9422f8f5516ac55971b6def5fe58ab1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4108,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=101;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4108,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d063ed4aa47d00f7cf72","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ba690f6c10cbd4c90ac1005","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae80ffefe9422f8f5516ac55971b6def5fe58ab1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4019,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=12;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4019,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-590262fd2dfce9f1ed44","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dbce9afd188a70e401a9d340","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7673bfecf276a4222926f6a0d939a0ae3175ea1f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4033,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=26;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4033,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c328303e39f8ec91b475","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3391ac7ae7a8fb58245c939b","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9972222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7673bfecf276a4222926f6a0d939a0ae3175ea1f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4106,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=99;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4106,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":99.72222222222223,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3ecde2db8bf3b26261ee","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b128300b701e36bf0d433f1","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7673bfecf276a4222926f6a0d939a0ae3175ea1f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4023,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=16;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4023,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.333333333333336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f909c5ef645607ef0cc","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-90e98b8a71a4d27b52f313b4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e5c49f61b2fb4724a0f21635eb1c6382d6f0cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4109,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=102;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4109,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ac65000733fb43b443e","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd030b489b3d9cfc98043d8a","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9777777777777777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7eff7d2a8e80893773bfdfb358caa51a41ba570","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4014,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4014,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.77777777777777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-698193d92c1f2f14d8d0","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-13b562b27604e4751a0c45d8","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9916666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7eff7d2a8e80893773bfdfb358caa51a41ba570","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4013,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4013,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":99.16666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e1b07b58c7d6dff59ed8","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ecc0809b8f799db7e9615026","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9611111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a9fd0c12a37311bda639edece75dd24c2c46a1a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4117,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=110;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4117,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":96.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4ec300a9eefc83fe6208","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-399c54d737e634200c6c0254","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f596f88092a44b7fbb4f28ce632c7d2ce549ec7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4032,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4032,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-69bee5d6395f542cf7c4","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-694898711bdff189d3443d30","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd4451fc689b102e64716c9712967ce521a2b20e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4022,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4022,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-538635f20e33e12744fd","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e7e099e68743e95ae97b5da0","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd4451fc689b102e64716c9712967ce521a2b20e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4021,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=14;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4021,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d7922cce090918959cbc","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fce64a24aaaa409eee2f865b","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9722222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd4451fc689b102e64716c9712967ce521a2b20e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4089,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=82;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4089,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":97.22222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d227e1c52fe40d303c8f","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d007c49245b9a5d548578d5","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"611afa1195130a97c715672f041d28a7b052de58","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4039,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4039,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1954280d53ac9ef1474b","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be5bb34b23c7647666a266f0","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b_none","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8222222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"611afa1195130a97c715672f041d28a7b052de58","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4040,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4040,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.22222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0b5773c0a5eaa5f4ed7","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6dd93a1432f838d6853cc526","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e4fb60da12c6c7465b790e54877be7894e601f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4054,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=47;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4054,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f87b349e8be69505d28a","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-62a86ddb7a506a93a1ab8598","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1e54308ab07ce7e859682f0c8f3349dd30754942","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4041,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4041,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-54cdf30ec881faed7993","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c5e8ac747958e3b58545e323","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026-04-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"775b066e298b478b382a8e7ca9bcb3c12a791ec5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4037,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4037,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bfcd9a455b4d6b6694c1","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e45d564b054731b6d062bba9","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b_none","observedAt":"2026-04-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6666666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"775b066e298b478b382a8e7ca9bcb3c12a791ec5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4038,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4038,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.66666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2dd06fe8c35adbe690d7","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f43413a03173147865216cd9","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f34af4ca68e6324896d2383aea7cbbe2ea47f54c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4044,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=37;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4044,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cddf8c18325866cb35a4","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-054aa2b5315890d555728ea0","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b_none","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6888888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f34af4ca68e6324896d2383aea7cbbe2ea47f54c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4045,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=38;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4045,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b15debe6327547e09d7","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6ef297144f96c91964a9bbe2","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5277b961d7c706a07b917377a64b7d42b21e16f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4051,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=44;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4051,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3c06ce64b9d5e7ec254f","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cd79bb0560d02e681c591ef0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9111111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ebd4205b4bdac80d31d00e4879b0d681cecdcaa3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4025,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4025,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.11111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e59e4778669852552e7b","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a894c155cfc85f2e7516e20e","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f11bbf85087ffd6ade22c14f018b65a2827afc0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4035,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4035,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-93613bf2a4153565e751","canonicalModelId":"qwen/qwen3.7-flash@2026-07-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c7d6dbea04100cc1e219c926","mappingCandidates":["qwen/qwen3.7-flash@2026-07-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Flash","modelRef":"qwen3.7-flash","observedAt":"2026-07-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8666666666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8f62b748e327beb6e9180d04cbd2cfabe2093243","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4048,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4048,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-07-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.66666666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-07-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d0e7627a59d9c2ea6f3a","canonicalModelId":"qwen/qwen3.7-flash@2026-07-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db633fbd156f5289864c9e8e","mappingCandidates":["qwen/qwen3.7-flash@2026-07-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Flash","modelRef":"qwen3.7-flash_none","observedAt":"2026-07-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7777777777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8f62b748e327beb6e9180d04cbd2cfabe2093243","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4049,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4049,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-07-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.77777777777779,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1ed820d5fc1aae7156eb","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-22b047319331640a35fd0cbb","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2345bae80457b4f85b11bcadf75ac19ec1f2762","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4028,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=21;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4028,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4fb7df8d26e1cd782747","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-54c1f6677f76521584670e9c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9333333333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c23ed870c408537a7c6ff3580d40fe03d63859b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4046,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=39;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4046,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.33333333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b7229e142ffb790eb5b","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dc6731c3d06cccab80af901a","mappingCandidates":["qwen/qwen3.7-plus@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.7 Plus.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.7 Plus.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus_none","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c23ed870c408537a7c6ff3580d40fe03d63859b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4047,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=40;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4047,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-otis_mock_aime_2024_2025","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-26cf744cc8348997c09e","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-05408cbe684e639d7f2a3057","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9944444444444445","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eaac5dbfad74ba73b6b7a902955ee8c1f66b05d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4082,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"otis_mock_aime_2024_2025.csv:row=75;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4082,"metadata":{"epoch_file":"otis_mock_aime_2024_2025.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":99.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-959b91c26955896373d6","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4ec6b0216068cf21987dab45","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2db9b222ca90d369c6561e7bd437bea398274aed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4415,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=55;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4415,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b9b74ba86c0282c8e861","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8674c2b6e0b537761bde5d9d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.03","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e872c1a9ed4bec9c1939bfed9460cc92806dc3ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4416,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=56;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4416,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":3.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7ad2be0ec37db0220649","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-512617e1a7c5197adae457c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4eaa42328a4972a2a54d18f0b65603603b28bf23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4394,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=34;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4394,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e6fd6fff62847d54a9b5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e97fcb3138af24d398cb8ad1","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d3c002abd0e063f15efe60fc5d404692e4fb8a34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4363,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=3;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4363,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ba19b2578b43ec9bfde0","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e10add905c59b6ac89008d7a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7aeb590c9a008e4be4cbf1adc4270d394aae50ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4382,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=22;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4382,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-08535df74fe4b81ca8eb","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-60e462b5b1abe0cb153e7f44","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3de9eae0658aeeefbcbb0d589136ab1251f8569","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4377,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=17;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4377,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b10e6b94bd0eb0cc31df","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b7aa0140236231a726842dc4","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.54","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e7d6dac997676197b3cac12d789b11ef9c3c5902","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4374,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=14;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4374,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-00a065ad1a45e607635a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4cb9977a0ad1bdb61d274d82","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6900000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf19554f277999f654701226e04e178834517162","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4368,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=8;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4368,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":69.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fcf0c3463eb4dace99fb","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08cb5e8af20a4bf7ab7cb49f","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69aee1e5b3551ed814dfb0c1e09d5be3e591265d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4362,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=2;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4362,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e9a7571d55e4faff099d","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-58cfc93cf5ec29cb12feede4","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3ef6a390890b2f96c6e07dcabff798f60ff8b14a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4393,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=33;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4393,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7c2412baf1056c8c812c","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0b251da9762836f5e1d8cafb","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24b0bea2da69e61a2ee81fb5875286b0e2d38dd1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4379,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=19;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4379,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a4c621b086d94b2ca146","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df405c3a103ab89c8dac4d18","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.77","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da7f425ee5e8b7625d4e6a4c8d029c7da4843d3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4366,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=6;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4366,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-58fa2cd52c18d99ffb41","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f3bac702baf74d4e4bfff126","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_unknown","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.56","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de17707d51b18a7204508d6aba44e1228872fb61","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4373,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=13;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4373,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.00000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-02b6e09cd497c2a8a64e","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-99833c2a62a070b6c1341942","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_unknown","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81ddc88286a5cce0d575963eea27f3cf417f1aca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4378,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=18;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4378,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-483245c7ecb8d0bd6fab","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-24a7011ea18374c5adabdcab","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c2c7b723f9b76d1c1e36d24689e40cad26ded59","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4399,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=39;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4399,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-71cbe0e4b73906f916f5","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c501734a6ba9999134d2ea36","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of DeepSeek V3.2.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of DeepSeek V3.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V3.2","modelRef":"fireworks/deepseek-v3p2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b832e0552aea560a1200e6824730147f79d326b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4410,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=50;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4410,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-89abb64907674573ef87","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-37997489cdaf35c0022efbcd","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8fbf9a8efbcf7e2f5993a7f523ce5a83ecdb37af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4388,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=28;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4388,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-86dedaf9f27ecbf6da05","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4ce248f17242102156eca0d0","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68d0b0968401f0129ef2c88177dbc5633b98c05a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4404,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=44;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4404,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b5f5bdc40ceff7b1112f","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6bb3bda4c2ea6b1a97533d9e","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd62eae3cb7a903bb9cb2e30edd2addea5745dba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4386,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=26;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4386,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cff4269cdab32e979a31","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-75c03890f4b8b04aaf60d4b7","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_unknown","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa7edcf520784d6a33cab6f6a81f3a1cca8f9eee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4383,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=23;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4383,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e438658c47fb6231f91","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-56a61a310be1afa658ffcdb4","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_unknown","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7185d45a5c55df8a8f9c89a930630e5f8334aeba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4370,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=10;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4370,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0060067f045f7380e36a","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-768fa09a655339e98628655d","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.06","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b7cd9e45a2e584e9213afe9846f95d2ba5bcc5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4411,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=51;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4411,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-16de1b1241f7612216a0","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ce28d963ab855a69e3bf3e9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.22222000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc7ce8ec8f60a0e63e8c78318df0dd46eae0889e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4389,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=29;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4389,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77b1cae7a98787132af1","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3ad13f3dad891ca9df217233","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.35000000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de97dfa7eec4dacb635c0b675faeb0f8e35d2f69","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4384,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=24;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4384,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a28da686d95109382ef0","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ef34e011f55f9aefffa1dd17","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.18","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59fa1fe2af7ce9f49257e742858e76b58fa26bda","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4395,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=35;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4395,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":18.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f8539a683592e08538ed","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32860098046de98e40ca30c0","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.09","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0071ceeddb8a5f6c15be7c29c8fb460649b36c8f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4407,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=47;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4407,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c762fbab65971444901f","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa5549e3efd33b6ce01746fd","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f6054e2e007332deba4c5d75a3bba22b75f9521","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4405,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=45;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4405,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eee38eabdc9d3879c5d9","canonicalModelId":"openai/gpt-5.1-codex-max@2025-11-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b4f1674996b22db57ba5614","mappingCandidates":["openai/gpt-5.1-codex-max@2025-11-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Max release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Max release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1 Codex Max","modelRef":"gpt-5.1-codex-max","observedAt":"2025-11-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.09","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60ce0c0cc3c2538f6bb105c3d3420445e1fd176c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4408,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=48;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4408,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2025-11-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8007d918aa286133a649","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-387a59129d3f18a25a5a3317","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.15","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1e02078d18397965fcb23d7cce058178d5ff71b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4402,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=42;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4402,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4130119b25ecf7c61071","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4688e8ccf49f5c78711797fc","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.56","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7244fc2b4897ffc01e41d748cd0bd7048fdfda15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4372,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=12;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4372,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.00000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dab14fd568b0ff9377ba","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fcc36daa52c66917db3c043d","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd3cd0e245602d485850c4315added98d0024faa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4391,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=31;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4391,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-30f333abf31036fb9570","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6cc9bacdc5355e5b27cb5ea","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54522a67e47fa295019f3773179806e12c38369b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4413,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=53;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4413,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aea8cbd25f86f4c46f7d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eac83511378a5e9c44f27eec","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"235e66d2eca63a12e6e50313410ccaf3e9f1269c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4376,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=16;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4376,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c4fe3258d4333b66810a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dbca2ae7fcbb852b9a53783e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f3db81a17ce524d78c697ddb897757fad256d00b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4369,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=9;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4369,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3a0f6c21298f29acb2a4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c9bcf6058572ed75ecb1ca28","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8300000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2c47fc720ad8410e1284a625431879ed874dd502","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4365,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=5;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4365,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a498064e96b05e0a57b4","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4d813e5664865f067d80d89d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.74","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ff9c6cb51715863d0b3f9e8f3b0c54456b03939b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4367,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=7;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4367,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0147bd4b66edf198bc8e","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-377814eaf02a6bd75d1867c2","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.31","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dedf7a12b6923167ec375c7f4c0bbb578ddbc2d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4385,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=25;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4385,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ee512c115b87796f7e1","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4d957ce1e33792febd47350f","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.51","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d81fe2294f3530091bbc745292a786bcba20d56a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4375,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=15;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4375,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":51.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0701357924d8d5297984","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c9c2b81af41927c9e190e49","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8082337b1e84a1f9c206a93fd1474bd53bebc14","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4397,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=37;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4397,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b135e70ddaa2f7701d4e","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-987ea620e694f5a2f7b1c65c","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcc156bb3199676f24971a34e4591ebb81d75926","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4364,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=4;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4364,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-59bf466e84927f2b8763","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30df61bba3d14af4e1a2d3c9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55851fbd65558b07ecd63b4ec147b55343981e24","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4398,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=38;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4398,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-c654c7fc33b0dd1c90c0","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-22f7fdba41669845335c06d6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent"],"rank":null,"rawValue":"0.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"903ae54838f2b8e83f8fea3bfb701996230ac2d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4390,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=30;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4390,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.1 re-grade (vals.ai, 2026-08-14): corrected Lean formalizations + hardened grader (rejects sorry/admit shortcuts, larger elaboration budgets). Not comparable to v1.0-era values. Updated 2026-08-18 after multi-agent verification.","organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8a0b874720b4ad247499","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2f01fb23e1258d78e4a27a0e","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a13a64bee3a32ccac8604003292512cc3fbf423f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4387,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=27;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4387,"metadata":{"epoch_file":"proofbench_external.csv","notes":"ProofBench v1.0-era value; model absent from the v1.1 board (vals.ai re-grade 2026-08-14, methodology changed) and not re-run. Not comparable to v1.1 rows. Annotated 2026-08-18.","organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-848d0ca1c17b365d3e93","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-731ed380ec13a281507dcf9d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aef0eaff8dde9475983d76adcc7fd5dc882003d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4400,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=40;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4400,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-proofbench_external","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cd0f6ea8a52b56a1620b","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7be97d382a6edf94033b926d","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.8 Max.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.8 Max.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_unknown","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4db9a8f17c71355e50b56127d0bd61f3c0d98da2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4371,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"proofbench_external.csv:row=11;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4371,"metadata":{"epoch_file":"proofbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-56e371f388d1da7b11b0","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5df68ad1cbe2919308bd0d59","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.161","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"207a54ff27acc9fd6b5bcd486d6079522f297b72","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4421,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4421,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.161,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-29fc174a6824457e30ea","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b43fda2c6335e1031eb14380","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.158","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"207a54ff27acc9fd6b5bcd486d6079522f297b72","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4422,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4422,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7ea02de201e7c7a31334","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d80275c82402326edab6c0ec","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49b2a0eb65560d60ba0d2adae9a2be2f655d8313","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4427,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4427,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d473f14e4757e93a260d","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ca7ead15c905342e259b552a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4105dceaf73a765efb8e3afc62c126e093425adf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4426,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4426,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9739ad51657ca2e90177","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f38db8e873eba2389f842deb","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0833","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d54fb644c91e9380be75f30ad541d162d241f83","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4423,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4423,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-06faaccdb28662d182ec","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-23488cd0590347ccbfe3c0bd","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.020800000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"606c5c88c2064b54c05dff42e192fbd2a428fe6b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4430,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4430,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.020800000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0529998898050e8cc797","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9764ef857872ee056150b114","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_unknown","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.05","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c141cc96ff92919f68ac9429dbf028ad442b0a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4425,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4425,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7a67a829c1f317e8f9f3","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3db575785bc64c819218207d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0167","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8e5bbac59e37b02a57883f4afaa54c59a4a4436","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4431,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4431,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-55dadb28b551577bb702","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db5fc514c2a98bde04323a97","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.025019999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d9051064efb636bd0d8de0e34849cfb08c5823f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4428,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4428,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.025019999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa60403a23c222802fd9","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0a15bdd2489f6b5ed176cf7c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.020800000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d9051064efb636bd0d8de0e34849cfb08c5823f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4429,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4429,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.020800000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-rli_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-79b109a00c8492cb8c26","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-356a912c3fc8d24b9b11233f","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.0625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7101904f39be7d038daac9b15693f1a90543da61","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4424,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"rli_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4424,"metadata":{"epoch_file":"rli_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.0625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-09-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a226a5c792c8e83a05f4","canonicalModelId":"deepseek/v3.1-terminus@2025-09-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-616b4c2fbd71bd23cd5a388d","mappingCandidates":["deepseek/v3.1-terminus@2025-09-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V3.1 Terminus","modelRef":"DeepSeek-V3.1-Terminus","observedAt":"2025-09-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.40625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24a1e4035eb08a8286b71c8138f7ec014852d8f7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4555,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=123;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4555,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.40625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4ed99e5401aa7e7a1762","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d3a3ea78f7d19c225de00738","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.388888888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8133a425dc861abec928b16f4da4beaddebe7bdd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4558,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=126;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4558,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.388888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-195655c90a1135e405d3","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-763461e4b0d5a502b434ec76","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.469907407407407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e026e764dcf6700e32ae96cba6906c1f5ce48d9b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4454,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=22;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4454,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.469907407407407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-83be185c8ed8cd64ca5b","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0a9970aca32608b44d90efa7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.45370370370370405","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9bb43f447ae6db5035300a776884240ab0623daa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4457,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=25;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4457,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.45370370370370405,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5b4ef3ba14400818c792","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4537ef42d26b9327ff19f705","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.601851851851852","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1438d0b909a840d7dd7106ef72144dc58a9c61a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4434,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4434,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.601851851851852,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-410c9ca1f7fd01c161f2","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0177a0a1f4f50b5178167151","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.43287037037037","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae4d227718b05a6e490bae16b6f38965267be2a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4461,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=29;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4461,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.43287037037037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-445a9b08f5a40159ca38","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c41567a6e2d968a9c9b18efd","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.545138888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d7faa3047de8290f684c41f1ef26ac01b9eaa70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4439,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4439,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.545138888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-929a4d49f6fe673ef982","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9a19f173cc1ff7b28cd1221","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.534722222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"507410ff650cfc5c5a79dd9df3aed0517d6d7a4c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4441,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4441,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.534722222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b517721c029773698f64","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c23e7c4dd574d3511a7f72ed","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.542824074074074","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"073738e982f443595bbebcbb4c4a2a72c742d5e8","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4506,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=74;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4506,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.542824074074074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d663b3447ce6b8d139ec","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9563121e91bc85ca3b1af3e","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.48032407407407407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"073738e982f443595bbebcbb4c4a2a72c742d5e8","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4508,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=76;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4508,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.48032407407407407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8c2ff1095aa80f1d47d1","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c87159f3ae2a28622ca3955d","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.556712962962963","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"073738e982f443595bbebcbb4c4a2a72c742d5e8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4504,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=72;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4504,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.556712962962963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fb5571369241cdef6eda","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2e14ccc76ae5c30e4a9e6842","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_medium","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.506944444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"073738e982f443595bbebcbb4c4a2a72c742d5e8","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4507,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=75;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4507,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.506944444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0c0f7792df8eb659ef19","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db6c17027247f8544db3a3b8","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_xhigh","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.549768518518518","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"073738e982f443595bbebcbb4c4a2a72c742d5e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4505,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=73;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4505,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.549768518518518,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-431566ace348958324c0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bfb95126f3949b769ae2feca","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.446759259259259","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"789ae958406272ad0500c976507407d325c47daa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4552,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=120;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4552,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.446759259259259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-272dcfc4c3d3b9d95125","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f31ac5a6ff5fc7b975a1155","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.467592592592593","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81c2f55cda38174e89b6c2cbff5cd6b3b7446bbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4456,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=24;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4456,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.467592592592593,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b090c632552cfcc5756f","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0cbff5a92553636711f475c1","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.486111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30ee295e932bbb5a066f958053c8a30d6acc0f59","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4546,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=114;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4546,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.486111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77a5d490967e8dd19828","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-23b28249669c066f0236e2a0","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.53587962962963","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30ee295e932bbb5a066f958053c8a30d6acc0f59","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4471,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=39;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4471,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.53587962962963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f5e15821a63321a1423","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0cd979c5f28c081adc2bd851","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.49884259259259306","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35cd2bd6d24bad655790d69b8aaaa12f34cc037d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4543,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=111;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4543,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.49884259259259306,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fbd03e5f38458c33a736","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eb6b6a8297942b2c83e39699","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.420138888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35cd2bd6d24bad655790d69b8aaaa12f34cc037d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4554,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=122;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4554,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.420138888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-41149d71c8408ecab3b7","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-79fb22233c738ca88addd7c5","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.44907407407407407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35cd2bd6d24bad655790d69b8aaaa12f34cc037d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4551,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=119;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4551,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.44907407407407407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-35d23cc077194120fff6","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4157a0ac33dab79fa29f536f","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.44907407407407407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35cd2bd6d24bad655790d69b8aaaa12f34cc037d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4458,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=26;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4458,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.44907407407407407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c61a94aaba7d7e557254","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c444af4809f98e72abf65342","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.46412037037037","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ed9e812dabb8e6eefd16b37685ea87fc898eace","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4548,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=116;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4548,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.46412037037037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-065ea92ab31c1552de07","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c10b0a6344757c6feaddd057","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ed9e812dabb8e6eefd16b37685ea87fc898eace","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4448,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=16;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4448,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-19bec1c005707a41e0a7","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6369c2f040f933ba9a27ef3e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.428240740740741","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ccbf5b68df2409cf8cf5254c0b368ccc3c9bb1c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4473,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=41;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4473,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.428240740740741,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cad24416a8cd0f072c5b","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-987b4314cf2762f76f8d5bc3","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.58912037037037","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"79c1e3bb8052aec59f686a9c309275362b00ce5b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4435,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4435,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.58912037037037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-528014e32697b6e83edc","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3d7f629e3e10caf9bb8ff6de","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.40856481481481505","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1553a65a0d794455dbbe38a8201ba6fcc94f7852","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4519,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=87;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4519,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.40856481481481505,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce9c440638aea9d3aac8","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-92d35b1fcb3657d91bbfae4d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.53125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"280f9c3b9a1c86e1aa9e87dfaf79cb2bea290d01","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4443,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4443,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.53125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f081d625e8359f8b6df6","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6dca6fe0cdf6bdb68e3e6bf","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.52662037037037","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf1d4ffa693579404bf03a6ce6e8aaf4e6ca7244","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4511,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=79;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4511,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.52662037037037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-16259c5e4f00fc938713","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3ca8883d3f05d8d2a273b53","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.451388888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e244b6c74bcad14d12004937722a969df9c6ff1e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4550,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=118;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4550,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.451388888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4521f5e00c215d3b7ed0","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cc00ad4fd9743b618f42d6e5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"15c2ae443345396beccbd71347411befaf6f9551","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4459,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=27;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4459,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0d986d2132396b2840ca","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3bdfd7aab1de3dd1a347a67b","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.50462962962963","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"997feeef352d50fccef0d4f13e57f34cdde332f3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4446,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=14;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4446,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.50462962962963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b8c5aa23cf27bdc1485b","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f930618e8eb878544b23b321","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.361111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"997feeef352d50fccef0d4f13e57f34cdde332f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4561,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=129;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4561,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.361111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e2c6fd13b67d957b0182","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7c6cf1ba6563087310d050aa","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.429398148148148","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8300c6618274a22f09543da092b2fb0edba9a158","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4502,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=70;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4502,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.429398148148148,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59b436e454ce9e181386","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-318dbd41f5a8ce782f5424ff","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.392361111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"67daaa295a3b88a8b42f2209368f2561d6cdc62f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4503,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=71;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4503,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.392361111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-67ed39689e27b61a8f36","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4cb2fe74a436860d27bba535","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.43287037037037","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea6c6413978863da37d24c9b932f609e793fda49","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4485,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=53;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4485,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.43287037037037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7463bc8827ebdc3db62c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7d90fa7d53f0777e021907a3","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.565972222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13f6ad85716db03ce319cdb238112e6357972bdd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4436,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4436,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.565972222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee0b43944f6593ee8ee5","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-027151496ae6d5438354666b","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.49884259259259306","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9865f27ef564cdc9283ae76a60e67356dc04af39","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4449,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=17;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4449,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.49884259259259306,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6de574c90947f0df2c7b","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-45116879f24886c1c090a746","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.46875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca1e6baddadba22ec6e5b1b9f54c461ea9b4a541","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4455,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=23;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4455,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.46875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-05-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6b4bc3db9d23548c66e3","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd7a22656f19e7866ecd8a3f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.486111111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1d57b664f0d2d522500efc6d439af2913f7429","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4472,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=40;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4472,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.486111111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-66822a0b8597fe6b59c8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-16ebec81af888f540d5a57a0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.559027777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1d57b664f0d2d522500efc6d439af2913f7429","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4438,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4438,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.559027777777778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-84acba8abab515375670","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-99e05655bb6b4f4d954c8cc9","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.516203703703704","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1d57b664f0d2d522500efc6d439af2913f7429","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4444,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4444,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.516203703703704,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-49d2b0d59421847cb90d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ffc26fca3cf6257220979588","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.534722222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1d57b664f0d2d522500efc6d439af2913f7429","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4440,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4440,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.534722222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f0121f4a4fb23a3384d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ca390ed0960b288fcf8d8375","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47337962962963","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1d57b664f0d2d522500efc6d439af2913f7429","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4452,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=20;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4452,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.47337962962963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3e9c9acbc13926224363","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c9c2448dc803829f02fc5eaf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.561342592592593","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f1d57b664f0d2d522500efc6d439af2913f7429","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4437,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4437,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.561342592592593,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7792c2cdb5d973e4c7ae","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9a97a1965455820c20ebb42","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.506944444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4212bda2090c97de83a1fb4d48d55b0253dd9784","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4495,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=63;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4495,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.506944444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc004c7cf8d1b5a4f232","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7dc8307fa0cfbeea3a79c0c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.456018518518519","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4212bda2090c97de83a1fb4d48d55b0253dd9784","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4501,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=69;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4501,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.456018518518519,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d82b1aaa0eb922a1002e","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b244001850b59ec525c82432","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.525462962962963","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4212bda2090c97de83a1fb4d48d55b0253dd9784","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4493,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=61;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4493,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.525462962962963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1ef9b9d9182c6ccde0bf","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0efb504b19f64914fe365836","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.45833333333333304","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4212bda2090c97de83a1fb4d48d55b0253dd9784","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4500,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=68;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4500,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.45833333333333304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-45148d68cc60a2323c0f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8e57ef996a14d2b9d6a54a9c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.399305555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4212bda2090c97de83a1fb4d48d55b0253dd9784","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4556,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=124;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4556,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.399305555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-136c50aa04e4c4245b77","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-af32d951b3c4ee9682eb799c","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4212bda2090c97de83a1fb4d48d55b0253dd9784","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4497,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=65;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4497,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4c2c67d986455a5de4fe","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-93c3ff1ebf2781046d1a18b8","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.569444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc1cd0b01cc9323ea7aff6aaab12b2b7b7920d97","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4486,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=54;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4486,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.569444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7dd2bf132115d42cee6e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e862afc99dac711001d9d2ad","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.554398148148148","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc1cd0b01cc9323ea7aff6aaab12b2b7b7920d97","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4490,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=58;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4490,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.554398148148148,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c7586113845df3a84a6d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-40a98bae325d28a744b644ea","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.561342592592593","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc1cd0b01cc9323ea7aff6aaab12b2b7b7920d97","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4488,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=56;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4488,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.561342592592593,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5c0b3bf56dba55444eab","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-75d3004fe62ca78e62d5e76b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.564814814814815","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc1cd0b01cc9323ea7aff6aaab12b2b7b7920d97","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4487,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=55;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4487,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.564814814814815,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-05c36979daf6aa49dcf4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6519d86647c43ebbe2a4e9b7","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47106481481481505","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc1cd0b01cc9323ea7aff6aaab12b2b7b7920d97","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4547,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=115;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4547,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.47106481481481505,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1af7051fbeff5efd4714","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f839632b8907bcf7396f875e","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.560185185185185","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc1cd0b01cc9323ea7aff6aaab12b2b7b7920d97","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4489,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=57;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4489,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.560185185185185,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ab0ad0c1345964100c3a","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0910929be26b898d650631d2","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.501157407407407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e541549aaf0d66fb38308778021c518cfa9d64b","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4496,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=64;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4496,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.501157407407407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3aad1f97cc6a169587f8","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-332ce289161761d15eca2353","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.491898148148148","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e541549aaf0d66fb38308778021c518cfa9d64b","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4499,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=67;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4499,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.491898148148148,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1109c56306171cc20e64","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-41499dbb654200d8bbe407af","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.539351851851852","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e541549aaf0d66fb38308778021c518cfa9d64b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4492,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=60;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4492,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.539351851851852,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3cec5aba053a771d4f5d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7f7aaf5c3e80cf9ca025f5a2","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.496527777777778","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e541549aaf0d66fb38308778021c518cfa9d64b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4498,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=66;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4498,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.496527777777778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4e33e992e59d1e6cb198","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9b607fd51d46581bdeb8813","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4456018518518521","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e541549aaf0d66fb38308778021c518cfa9d64b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4553,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=121;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4553,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4456018518518521,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3d3d7ae3acfa4974f852","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3b9e9a78dc083dc22f91f4a9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.516203703703704","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e541549aaf0d66fb38308778021c518cfa9d64b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4494,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=62;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4494,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.516203703703704,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca08e6b3ba4cdf425865","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8a2a0ee33e047dd257bd773d","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.540509259259259","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e74e3e71ba702fa1c42e6dc338b79a72cd55a3cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4491,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=59;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4491,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.540509259259259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9d0f6af17475bf2008f4","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-21b14429ef6c7ee3fa5893e3","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.48958333333333304","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5aad682ff855adc28ff4eef95d872e439c08e186","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4544,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=112;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4544,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.48958333333333304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-142ae47581463706d5bc","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4f9ce973267313e6c5887a8d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.534722222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c787d21b7aac13ea46d0224dc2c4ea7618b704e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4442,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4442,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.534722222222222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-eca475103037d1807ac9","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-10d700b25a4a586197be0784","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.474537037037037","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c78116c47d0605beec904efef92ec8a9782e0663","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4451,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=19;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4451,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.474537037037037,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a9c635e9ccd2f6f51ce","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bf15508df436876d594b38f2","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.511574074074074","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4a3e381ca2d66d7bb0afc2493a5844ae89b288f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4542,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=110;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4542,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.511574074074074,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-08f6102ead06e2c44e86","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33008d65d61ef036e16d4cca","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.586805555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4a3e381ca2d66d7bb0afc2493a5844ae89b288f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4541,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=109;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4541,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.586805555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-82bbd170ca5ba7e004b7","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ef6987a16ff9b3364c51acb","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.586805555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4a3e381ca2d66d7bb0afc2493a5844ae89b288f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4509,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=77;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4509,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.586805555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-93c83e8561a38a85972b","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cebe323335e47504a6c3c551","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_unverified"],"rank":null,"rawValue":"0.259259259259259","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a06e2aa35fd044bfe0c71a603ecfe4477fce60ac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4481,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=49;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4481,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.259259259259259,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-77825b93e3b149db2783","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c52760e0867b6b589b44f841","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["unit_unverified"],"rank":null,"rawValue":"0.430555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d56fcc2a47f0cf51a5a65aa33bd7aedc6f525ed0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4514,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=82;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4514,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.430555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-102033d088e6f27c940b","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2d5b9394aeacad3574c8aa0d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["unit_unverified"],"rank":null,"rawValue":"0.502314814814815","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a159c2a53256201bec282da4ba72d19073f9ff2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4447,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=15;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4447,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.502314814814815,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-02-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a5fb4e2d2428edc04136","canonicalModelId":"qwen/qwen3-coder-next@2026-02-03","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9d3f3cde6caa5e374830283b","mappingCandidates":["qwen/qwen3-coder-next@2026-02-03"],"mappingEvidence":{"confidence":"high","note":"Epoch exact Qwen3-Coder-Next spelling.","registry":"model_aliases.json"},"mappingNote":"Epoch exact Qwen3-Coder-Next spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3-Coder-Next","modelRef":"qwen3-coder-next","observedAt":"2026-02-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.322916666666667","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5916fa5694f3169cb95280f1b8e310a6f787d8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4531,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=99;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4531,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"","release_date":"2026-02-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.322916666666667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0afcec4e05c8e252407a","canonicalModelId":"qwen/qwen3.5-9b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-012188073aac34d35b101063","mappingCandidates":["qwen/qwen3.5-9b@current"],"mappingEvidence":{"confidence":"high","note":"Epoch exact capitalization for Qwen3.5 9B.","registry":"model_aliases.json"},"mappingNote":"Epoch exact capitalization for Qwen3.5 9B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.5 9B","modelRef":"qwen3.5-9B","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.275462962962963","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4c2b411080a9c1c2cf58dc6930b4829094898ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4479,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=47;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4479,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.275462962962963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-80eff6421b38c67fa930","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9a6170106f104b92772ef33","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.357638888888889","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b168e1f555f0e4fd061f6701239a602dcda310e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4467,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=35;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4467,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.357638888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ed136a0f19bce229f557","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-06fcd0284e4582bbfe808a25","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.407407407407407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80e2fa76ced57b8d6a57a5d813d143cd29fdcb19","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4463,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=31;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4463,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.407407407407407,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fb856fe951efc17a6700","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-af0bb3240c4f1f82a03e71dc","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.488425925925926","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8932fb3b77ee3fae685803833589b6c424ae7f91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4450,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=18;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4450,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.488425925925926,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-scicode_external","benchmarkName":null,"benchmarkVersion":"2026-06-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fbc2bc46c03eaa57372e","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-da5125d3d709c6828b9f2c35","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.454861111111111","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49bc69a886b066de999b6bc346b5da84e6a02132","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4513,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"scicode_external.csv:row=81;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4513,"metadata":{"epoch_file":"scicode_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.454861111111111,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-79dff3b603c01706e374","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9cf66f2bccbd7128419b84e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.408","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ff642982cd58fa7e77aa0e92020eb0040e5ebdd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4659,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=73;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4659,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.408,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-08-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-08f50f4808ada9762929","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e4106f72db561ed2d141772b","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"DeepSeek V3.1","modelRef":"DeepSeek-V3.1","observedAt":"2025-08-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"973f1e8d7b92ff8ac204ea3278bfcca6695229b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4661,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=75;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4661,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-08-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c3751090515a5d64869b","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7673f3ac4617acf783dc3b62","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.45799999999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f1c0617174036150313e17912f374d031de3971","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4652,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=66;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4652,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.45799999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-77241ddb147cbcb29489","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-db742de79d6482ba46b437e3","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8190000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"edda057977ec003cc48ab18ad10cdf43337c9199","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4589,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=2;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4589,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8190000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5ea14629e2db98d4b6f9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-802a920e95ea36d0c59e925a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_12K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.588","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4084f5fe25bcf1c55864b471bafb868ea15272c6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4622,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=35;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4622,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.588,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-22956d1f75cc43150b59","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5b9c6b5369d3a7baad94e59b","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.588","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4084f5fe25bcf1c55864b471bafb868ea15272c6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4623,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=36;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4623,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.588,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-910a845d12f1b9aa2cc5","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6b7af66850e6835a08bb8ec2","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.62","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a5322b490f24b4640e20a98fd24280a7e063c2ac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4610,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=23;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4610,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.62,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e1bb775accacc83bed3d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a21bd23887ea5f4a95f9b64","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.62","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a5322b490f24b4640e20a98fd24280a7e063c2ac","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4611,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=24;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4611,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.62,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-64477956084bafe4044a","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e58df336e36cbb20897a0ee4","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6759999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f684fd95c914d0c48dd9bd02406a9860806ea509","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4604,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=17;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4604,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6759999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-162366a482a44d7955ec","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f0ab10d4ca225b56457ce118","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.617","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36d2a86baf32a0d668d526896bae97388e45ce27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4612,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=25;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4612,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9b8c9d4be055c87506c5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5bc2feeaf56135113ecc56f2","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.648","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"15ae1424478aa3f2e7219108433f0e08d6d9e270","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4605,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=18;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4605,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2111f8e0917909f07e95","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b57dfdcfbcea1eba9b77282","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 5.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Opus 5","modelRef":"claude-opus-5_unknown","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.8059999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc512f801f4684bac29d8bf7151c851d646c74de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4590,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=3;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4590,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.8059999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-764be8e466f259dd6a54","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0501cf34b8116ce826e3f4bb","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_12K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.543","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"230e7b17478ceb2285cc137f5a43c5ccbcd0cc66","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4631,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=44;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4631,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.543,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-722ff0e0ed6a711b3c6d","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba7ee41c4e58f6f0329b8727","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.543","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"230e7b17478ceb2285cc137f5a43c5ccbcd0cc66","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4632,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=45;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4632,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.543,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a086bad4be435c6c3d4","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d1004bd1a3758e2d931aa274","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.606","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3eea71f2730cec0431fbc9d7b1977d1265e80d26","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4618,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=31;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4618,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.606,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6ce6a6112f520c88fed7","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d3b6dfd0e85414e19071a065","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_unknown","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.611","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a038f608fe38435ee25cdbbb5d15e19d4b1ac000","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4616,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=29;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4616,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.611,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-81a50c6a95dda6ce6318","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b1b02207cd2d54292343d9c8","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffixes on the dated DeepSeek V4 Flash release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_unknown","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.46299999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a038f608fe38435ee25cdbbb5d15e19d4b1ac000","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4649,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=63;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4649,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.46299999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a0df2cdae7837c5b6e6","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7febb86b620e8676f56c7cb5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.412","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d693667c9ce4f43b94943446bff7b064b214a8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4658,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=72;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4658,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.412,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04d897ba827371d5411e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-26c2f0f653c7ca82504e4fcb","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7959999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b573ed6376a6967c9197e04bf816c6d8939e4ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4591,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=4;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4591,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7959999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dab495f45c08dd97c4a7","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ffe698f82f9f55cac92c8a3a","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.767","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bb0663dd3b4b12a630f567f4478ed53dcf808574","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4594,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=7;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4594,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-51a2816093093eed5834","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4158a519a8aa708bcddab697","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.477","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42d93b00be5a2d9dd99a9e2d95c7d51107c13ba0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4643,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=57;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4643,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.477,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4358e27f905ca69999c0","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-639b4bd168029c488219a104","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.532","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85751e7e05785919535e9d03445b3993372aa3da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4634,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=47;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4634,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.532,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-80c758d29fb8757e9a89","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a7d91f5f186d5f99bb84eec2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.551","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23cb056075418deec068c44e5014a6a4d4e0034c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4630,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=43;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4630,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.551,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-afe3274fe5dc91013ef5","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a5fc187883b333a4413b9080","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.588","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3106cd11791aba0d24a14fe43f83d731a446fc92","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4624,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=37;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4624,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.588,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dad91a7550494fe6f521","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6419ac69436044e056b5fc91","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5670000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0275790e3227a2d8d2072b85ee54ca04bba54da4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4627,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=40;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4627,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.5670000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6486823e4596546dff8e","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed76eb76f23627038512304f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.532","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a7a6f6349ce7848a68cfea3f89dc3b2e57f4c2a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4633,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=46;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4633,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.532,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-94e93ac0cd05d9e72f59","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f2e03803274e33f813396aa1","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.458","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e8a514593dcdb20c22bee8b4029c14360ebc233","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4650,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=64;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4650,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.458,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-499470a4881ccee561b8","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-68e156028ba3347791c4e72c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.45799999999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e8a514593dcdb20c22bee8b4029c14360ebc233","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4651,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=65;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4651,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.45799999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-461acfeb6473e0343a3a","canonicalModelId":"openai/gpt-5.2-pro@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a930b31d0d13c4c1ff8d0b68","mappingCandidates":["openai/gpt-5.2-pro@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Pro serving tier.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Pro serving tier.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.2 Pro","modelRef":"gpt-5.2-pro-2025-12-11","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.574","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"341d3e6edeaa9bc0bb266f41066919d2d517762d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4626,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=39;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4626,"metadata":{"epoch_file":"simplebench_external.csv","notes":"5.2 Pro xHigh\n","organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.574,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cdf9301be0946fa0f592","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0e1f71882a15a46cc5002c8d","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.741","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f83cb7ae5e5afa08744f2a38a8f10ad50b16ec5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4598,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=11;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4598,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.741,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-637dc9f40b172a5a28f7","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2966eaa566be6508c8ee8dc1","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7690000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d9a9180ea47f9a16730665c494bba941f57470c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4593,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=6;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4593,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7690000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-efc6f93a46a1a3530fb2","canonicalModelId":"openai/gpt-5.5-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d9266c96de3686f093f64966","mappingCandidates":["openai/gpt-5.5-pro@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix on the distinct GPT-5.5 Pro endpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix on the distinct GPT-5.5 Pro endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.5 Pro","modelRef":"gpt-5.5-pro_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7690000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d9a9180ea47f9a16730665c494bba941f57470c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4592,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=5;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4592,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7690000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f0adefebcccdffa0ee40","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c21c637ec3b23541e656ed36","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6900000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae2bfd937524638dd8f66582f5da831bd2338129","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4603,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=16;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4603,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6900000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fd55d9a984bb47bc7e0a","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0c6914561641a12491e5627","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.468","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50f8e6c9901f8244aecfea7b8d62e259fc5c5558","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4644,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=58;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4644,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cd65995399de57b51460","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0d595be343ced48d4a7bcc13","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.46799999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50f8e6c9901f8244aecfea7b8d62e259fc5c5558","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4646,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=60;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4646,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.46799999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-33740492d7bcf61853ed","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-973939c83c348398a527b8f4","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_prounknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.717","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3324ad7d2fb29ffb8ab4d4dd79bc520ab148ec0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4600,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=13;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4600,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.717,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e9cb42888af0f5579f0a","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-72072a9bba5f2fbcfb54ca82","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_proxhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7170000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3324ad7d2fb29ffb8ab4d4dd79bc520ab148ec0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4599,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=12;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4599,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7170000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6deabb20f055bdb100ef","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d36330ef54066f3f47f0ecd4","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.648","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3324ad7d2fb29ffb8ab4d4dd79bc520ab148ec0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4606,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=19;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4606,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2571ad6805826b87ce9d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3f946af8ea3314044e978cc1","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.648","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3324ad7d2fb29ffb8ab4d4dd79bc520ab148ec0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4607,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=20;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4607,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e9f6650b26c0fcfd3a0d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-11b6bac5ee7a06778589b58d","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.489","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9cba3ff50fa26d45e4b0764bec6e2ca950a7d24","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4640,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=54;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4640,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.489,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9eebe3142fbc745b85f3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6f0a5c7c9f45ee58e6060e45","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.489","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9cba3ff50fa26d45e4b0764bec6e2ca950a7d24","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4641,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=55;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4641,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.489,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bdad1f527824a1fcb965","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3f0950c7b2bfa6a3dd684f0","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7000000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fcb6bd32b080d7cf850cfd7eb1da5dd7f3a8814e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4602,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=15;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4602,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0036923be45b11bf4586","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86e6ad2ac3cba46b50a78505","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7590000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0282e6b2bc94fc6a53978c6d0a31fc2a66e5c40c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4596,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=9;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4596,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7590000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bce617751aaf17a1f131","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b6d7cf94e0dfc0f24869a91c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.46799999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e44310fd0078488d3f875222e180e5a91adc4e74","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4645,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=59;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4645,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.46799999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d2598e3ef31dc1ce9193","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9e93b1a8b2792299d95dbed3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.579","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"282a37a3ccc059c6e65e74f9f052d2fa0d630853","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4625,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=38;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4625,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.579,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4762a6ab5bcf99908c25","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9543beb289a09ec4bd3b0b34","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.6070000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9cce7e7f0451f74d1313269eca7cd6b5b4704b2a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4617,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=30;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4617,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.6070000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-17116e7b955d53c91ba3","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cdf6df1cdb018dc49a431c79","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.352","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e1c23099edf089d008fb4027569430cb68ecbe0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4665,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=79;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4665,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.352,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-476a1f9c065b71a9c188","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f27ec28fa512dac37c3dcec5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9465bf0ac9e2d617171d3e60c19073313833bd23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4608,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=21;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4608,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-18e1f424e5857e7d0222","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1de0fb3e23880be31acfa7b0","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.7040000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a536c4c699bdaa693104c98c56875cf515bc74f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4601,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=14;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4601,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.7040000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-simplebench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-167e6934e724b3f7f75f","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-baa1225e286dd2672d7ea726","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score (AVG@5)","modelName":"GLM-4.7","modelRef":"zai-org/glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.47700000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42d93b00be5a2d9dd99a9e2d95c7d51107c13ba0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4642,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simplebench_external.csv:row=56;column=Score (AVG@5)","sourceMetricId":"Score (AVG@5)","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4642,"metadata":{"epoch_file":"simplebench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"fraction","value":0.47700000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-spatialviz_bench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2877d3853750123f4f93","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-519190f15c2a5d22ac52b26c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3686","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"64daae29eaad3aa61ab6ed223b510e055f7c9d60","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4775,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"spatialviz_bench_external.csv:row=7;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4775,"metadata":{"epoch_file":"spatialviz_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-spatialviz_bench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75aeec7f921db76509f3","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2383773e29fd156553f4c329","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Overall score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4466","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a85afc66aa9cf7aaa2ea2310bba483c9635c0c1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4776,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"spatialviz_bench_external.csv:row=8;column=Overall score","sourceMetricId":"Overall score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4776,"metadata":{"epoch_file":"spatialviz_bench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.66,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-56117952946033e723b4","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-749a86d3ff20907fd4899477","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5499999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f58a50120475591ea289b5bc8eed1c0b66212124","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4801,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=15;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4801,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1200e183245e7c446940","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8750e373dfb50d073af5b216","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc2ec13cbf3e37a1dd51e4f26a9be2c7771c79fe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4788,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=2;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4788,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-81ef62e4568db91eebb4","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3db1d98e49fb0d5077f2039c","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bac0c38d22665bacaa4f55d3b4b71cca5b5f915","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4792,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=6;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4792,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-51510b672f50755593ea","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bfe1bc146ccc8875d84ca503","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.68125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bac0c38d22665bacaa4f55d3b4b71cca5b5f915","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4798,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=12;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4798,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8d59e991d52f3bd4f080","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3f182f965584a20448a8907d","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_medium","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffc3bb714c5300dbf7a621449d59e7a204161005","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4807,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=21;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4807,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0fd79c97cd5ecf9fc150","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-572b0de38e289df05551f606","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f02141ef628cf5b888f34af6b38ad30a84438276","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4799,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=13;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4799,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a70224375f3dafa05ace","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3d50444978576bdb89954dbd","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_medium","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.58125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9706894d4eeddddb78387c8b0b51bd5ed0f66b6b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4806,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=20;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4806,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.12500000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7496eec9df6cf0e88c55","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ca32a9b777d1afbad5404d39","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GLM-5.2","modelRef":"glm-5.2_high","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.55625","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2977c2089f354994adfc0385137a454dde9c8104","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4797,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=11;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4797,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.625,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-efc51339cbeecd04b527","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac2a3a419961d253b40aa14f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.88125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1245b3edc9b2e990ddf56e5a5596d5575656ead0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4790,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=4;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4790,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-41351e122afba157e706","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-32910fdf1908134b5ac606e6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5.5","modelRef":"gpt-5.5_medium","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1245b3edc9b2e990ddf56e5a5596d5575656ead0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4794,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=8;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4794,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":81.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5298809a226dff967caa","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f8df3591d07cc9ef171edd2d","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_medium","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.61875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a8b9bd3a07f8343a396d5e23d135f32258be2fd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4800,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=14;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4800,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":61.875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-91ca47c40d50053d8982","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dd540ca3130e49a8ff3e6c7d","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.93125","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"08cd170c029341c3afb0345e0c91e85925ae2176","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4789,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=3;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4789,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.125,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9ffecd11fe0a5f60e5c8","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-82809c08bc79f4966b0eafc9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3426f7b56f2daa109606af0e36703406a0d9eb91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4793,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=7;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4793,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0f3e7278d023a546081a","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9f7dbb65315c88bb5b934a99","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.74375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea3bfa69251f59e151a020c91cf6c77fbbea677d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4795,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=9;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4795,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1b28a6a9060b3091a864","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f6701530ff19e079689ba8b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4875","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd9158635918ac3c385e5bd74a0837943fefb0ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4796,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=10;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4796,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0c3cc0d47160afbb3596","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-225e9cdc9d12423c20dd3e9f","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d5b1eb3cccd840b496f5308a071ed17f0a60eec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4791,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=5;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4791,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e87fef1dc840ee3a004b","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-00898e199dd776653ca5eddf","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.9500000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d5b1eb3cccd840b496f5308a071ed17f0a60eec","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4805,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=19;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4805,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":95.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04cb147aa1b4013bd5a0","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3b2ad2ecce88cf7ed865fba","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.44375000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1a06744bc1161c36a23cbd432be6be76eb0279a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4802,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=16;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4802,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.375,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-surface_evolver_bench_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-45fcfa1e3a5d663f4e8d","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-69f86c60df127c84f1f77860","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Mean score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b_xhigh","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.45","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"105a1213547df1bc190ac7a2bc012b002572e30e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4812,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"surface_evolver_bench_external.csv:row=26;column=Mean score","sourceMetricId":"Mean score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4812,"metadata":{"epoch_file":"surface_evolver_bench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e90c8eff3c9f53aefbc0","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-4d5830019e6685cb2f714d42","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"MiniMax M2.1","modelRef":"MiniMax-M2.1","observedAt":"2025-12-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.292","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fdcdf6e4ce0b3d1d16f9e016e9ff8309e02561ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4984,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=140;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4984,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2025-12-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":29.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2fe98d401a5081a24878","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Crux","harnessId":"Crux","httpStatus":200,"id":"pub-41d4f76695e499186a087c63","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"MiniMax M2.1","modelRef":"MiniMax-M2.1","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Crux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.36600000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fdcdf6e4ce0b3d1d16f9e016e9ff8309e02561ab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4960,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=115;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4960,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2025-12-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":36.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-be3b4735041cc019b14a","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"cchuter","harnessId":"cchuter","httpStatus":200,"id":"pub-c2559f3d0d1fb4b8ada4c66f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-03-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"cchuter","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.42700000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0307569cf1e42fff97f2678c332baf08468d8c91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4942,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=97;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4942,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":42.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa026e5b5a0dbaea4c6b","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-6f62f5dffd157418e706fff7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.42200000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0307569cf1e42fff97f2678c332baf08468d8c91","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4948,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=103;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4948,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":42.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-335ece61f48f9fc4c74b","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"IndusAGI Coding Agent","harnessId":"IndusAGI Coding Agent","httpStatus":200,"id":"pub-440acb875af18653e0dc1fc2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"IndusAGI Coding Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.451","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85fc242675f63ab56a20c212dae9acbd2c80ab70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4927,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=82;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4927,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=indusagi coding agent"],"subjectType":"system","unit":"percent","value":45.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6aba613ca0b88b35a72d","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Harness Agent","harnessId":"Harness Agent","httpStatus":200,"id":"pub-25b5cb9e37f2a50c7a413712","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Harness Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.429","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85fc242675f63ab56a20c212dae9acbd2c80ab70","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4938,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=93;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4938,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=harness agent"],"subjectType":"system","unit":"percent","value":42.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e358d56e15274bb7661d","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Dakou Agent","harnessId":"Dakou Agent","httpStatus":200,"id":"pub-c696103f75cd0ddb4cc74433","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-12-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Dakou Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.272","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9683409beb487e0d9c9c738d4dde46dc567a9177","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4994,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=150;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4994,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=dakou agent"],"subjectType":"system","unit":"percent","value":27.200000000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3028815215b65eca212f","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-5abb260787ee9e9509203e35","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.254","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9683409beb487e0d9c9c738d4dde46dc567a9177","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5002,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=158;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5002,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":25.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa4a2991cf3dc8e63cc3","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-54c3cebb93d4582b9180dda4","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-11-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.239","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9683409beb487e0d9c9c738d4dde46dc567a9177","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5009,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=165;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5009,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":23.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bab33cb2f92627e7b7ea","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-1f936642056d2fd5900a4bbf","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.254","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9683409beb487e0d9c9c738d4dde46dc567a9177","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5004,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=160;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5004,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":25.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a8282e8807fe5993693","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-09fa63b333558fac4d745177","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-11-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.239","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9683409beb487e0d9c9c738d4dde46dc567a9177","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5010,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=166;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5010,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":23.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2fd440d88f0e1c232568","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-c2bd7dda0cdf20d175b51ea5","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.139","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5035,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=191;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5035,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":13.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f328648eb90ad0fb75f0","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-bb47524979b739602c2c8ec5","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.283","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4987,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=143;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4987,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":28.299999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c08c1fd54bce08b84c5f","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-ab2f8fc672c975a3b584cc07","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.298","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4981,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=137;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4981,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":29.799999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3afc9cd142be147385dc","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-4c0aca0e57e0ace47ea7baff","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.275","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4990,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=146;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4990,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":27.500000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fa2a857a5f4110be600c","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-a708fad4f4ebeb48ad582b43","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.139","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5036,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=192;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5036,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":13.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59b554da8bc63017194f","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-8caf83156ff5088560f1be06","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.275","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4991,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=147;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4991,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":27.500000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e9dcdd1a3e6ab662561e","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Goose","harnessId":"Goose","httpStatus":200,"id":"pub-3fba274766f2654402fd81b8","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Goose","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.355","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4963,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=118;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4963,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":35.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32b3269d848b6fa9496b","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-2ca5300243b2cba45a8c5823","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.298","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4982,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=138;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4982,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":29.799999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a5e87dc25473424ceaf3","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-0de02ec4bfd66e2290021df6","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.28300000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658b2b183df694a92e15147fa3d9ee16fba01b29","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4986,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=142;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4986,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":28.300000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6cf69c48e5aec3d3a4ac","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-9d6f2c3278688c320abca4be","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.631","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4885,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=39;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4885,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":63.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b942a3b8a6b286ed8225","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Letta Code","harnessId":"Letta Code","httpStatus":200,"id":"pub-083007d9b54c2450c0b5f96a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_128K","observedAt":"2025-12-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Letta Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.591","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4899,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=54;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4899,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":59.099999999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f9327a78177703a9bff1","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-e0eafd8e9dfb241210c19247","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2026-01-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.519","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4916,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=71;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4916,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":51.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-80c95ed1a3564b88a2be","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mux","harnessId":"Mux","httpStatus":200,"id":"pub-d61e6e0a5a23d81db1b50976","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2026-01-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.584","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4900,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=55;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4900,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":58.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-21fc0fb50cd6fcee231a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenCode","harnessId":"OpenCode","httpStatus":200,"id":"pub-b5bf0c0d509909c871b0eea0","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2026-01-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenCode","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.517","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4918,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=73;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4918,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":51.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32f2e1cfa2e351213a04","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-aa4f79aea5df09e7ec745c07","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.631","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4886,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=40;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4886,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":63.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6d17d8f3e6d2ebc7e8fd","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-6515bf6e896e6135ddf84291","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-12-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.521","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4915,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=70;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4915,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":52.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f4eb63d07a2393c4d0eb","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Letta Code","harnessId":"Letta Code","httpStatus":200,"id":"pub-647043b88d2d156e7784cbdb","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-12-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Letta Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5910000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4898,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=53;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4898,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":59.10000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c667e4b86fda246821f5","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-3d6e23d4e87f7d3283bf559b","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.578","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4904,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=59;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4904,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":57.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d596b00a8b62822849f6","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Goose","harnessId":"Goose","httpStatus":200,"id":"pub-3617bcb040a06fbd52c0d5dd","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Goose","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.543","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d231f2cf3fd54be5fb4504626fb57b622eace10","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4909,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=64;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4909,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":54.300000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-572fdd07ca43889619fc","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-911c91830d9e4f1ba66b77fe","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.699","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4873,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=27;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4873,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":69.89999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c715b1f7b939450d8536","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-fee8dcb2af1ddde153600f9a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-06","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.629","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4889,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=43;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4889,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":62.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-23d4f51db82b24f122b9","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mux","harnessId":"Mux","httpStatus":200,"id":"pub-face9a8bfd517360dad1f49f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4877,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=31;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4877,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":66.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c066ef0fc99adc01396e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"TongAgents","harnessId":"TongAgents","httpStatus":200,"id":"pub-f72bc5b6b65b276ce4903692","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"TongAgents","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7190000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4869,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=23;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4869,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=tongagents"],"subjectType":"system","unit":"percent","value":71.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b6f8691e7bd0763a32f9","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-cda29aa343922d9b9940543d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6990000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4872,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=26;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4872,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":69.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e23f2e93e592d75ddf95","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus-KIRA","harnessId":"Terminus-KIRA","httpStatus":200,"id":"pub-beb3865bc0301fe7d434ca38","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus-KIRA","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.747","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4866,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=20;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4866,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus-kira"],"subjectType":"system","unit":"percent","value":74.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-40e312337d6b647ee6b6","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-b81cdedfa9d5ac7e905d1157","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4901,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=56;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4901,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":57.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-20ef4d8f9a34b2386e4a","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"MAYA-V2","harnessId":"MAYA-V2","httpStatus":200,"id":"pub-9163f767defd7063d6ebc79e","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-03-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"MAYA-V2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.721","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4868,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=22;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4868,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":72.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad2fe57f94806755b542","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"ForgeCode","harnessId":"ForgeCode","httpStatus":200,"id":"pub-874cc2dec96c7fa5dadfbc2c","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-03-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"ForgeCode","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.798","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4855,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=9;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4855,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":79.80000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3129d72a2e58e4a63022","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Crux","harnessId":"Crux","httpStatus":200,"id":"pub-65f07132a3fc7071ff4013f2","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Crux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.669","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4876,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=30;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4876,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":66.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e60c2c9d77e6e2b6894b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Capy","harnessId":"Capy","httpStatus":200,"id":"pub-5717243c51eb715f70af7e5c","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-03-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Capy","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.753","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4863,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=17;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4863,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":75.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c42144d31a7c44fb8036","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Meta-Harness","harnessId":"Meta-Harness","httpStatus":200,"id":"pub-4736554417b70ea13ceab7bf","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Meta-Harness","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7640000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d81721ecedc6751975d1fdb80949c2c1d6c3d63","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4860,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=14;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4860,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":76.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e8eb91b784a6ea28eb62","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"WOZCODE","harnessId":"WOZCODE","httpStatus":200,"id":"pub-e829b4b6d85e9c8b25544ac7","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"WOZCODE","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.802","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ec3c8579a54cd514ceec798762087386daf03f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4854,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=8;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4854,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":80.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-408cc73695257cc964e8","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-d5e55700f4450f7ab4746cee","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.428","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4939,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=94;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4939,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":42.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fceb25770b55b77e3b78","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-d561eba1e078e80796a228d5","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.401","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4951,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=106;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4951,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":40.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-85685a3d49365d7a6048","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-ac15a741396109eebdda7142","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.425","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4946,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=101;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4946,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":42.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3689d3c1f00395c24659","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-2a4c758bad5979d4721f8da7","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.426","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4945,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=100;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4945,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":42.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-038a8ea09b7ffcbeeaff","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-e99015ad7fac067d4c5684cb","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.42600000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4944,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=99;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4944,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":42.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59c6d250ed0b395e7678","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"MAYA","harnessId":"MAYA","httpStatus":200,"id":"pub-a42ed10bbd1e045add6d47b1","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2026-01-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"MAYA","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.42700000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4941,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=96;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4941,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":42.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-973f3db756d8b403ef62","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"CAMEL-AI","harnessId":"CAMEL-AI","httpStatus":200,"id":"pub-695d45d5f53c819af0f30de4","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-12-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"CAMEL-AI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.465","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4926,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=81;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4926,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":46.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c0c5b45553dc333c4e10","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Goose","harnessId":"Goose","httpStatus":200,"id":"pub-63fc8f204e6649f8c2fdd2b0","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Goose","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43100000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4936,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=91;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4936,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":43.10000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-24e9a0b98995eddcc745","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Claude Code","harnessId":"Claude Code","httpStatus":200,"id":"pub-5a5ac54e224a83a7291fbf2c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.401","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4952,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=107;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4952,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":40.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8a2ec8fc2f15a8355c62","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-373419f17758115a28269f4c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.428","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4940,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=95;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4940,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":42.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d5e475eeb8c50b593b9c","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-1849983e1d188d8b700e2677","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.425","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":11,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4947,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=102;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4947,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":42.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":12,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9eb5a01d623da067343b","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"MAYA-V2","harnessId":"MAYA-V2","httpStatus":200,"id":"pub-0bbb671a6eec61d21c8a4274","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2026-01-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"MAYA-V2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.42700000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2fe6055d3c6a00edd9eccc6fc09b379256cbb8a","selectionRank":12,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4943,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=98;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4943,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":42.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e4f3f19189391acd576","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Simplai Agent","harnessId":"Simplai Agent","httpStatus":200,"id":"pub-ad3d7110eb24f6d658a4779f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Simplai Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.534","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fa5d8744a4803e1b977f3e903b4a93fd643b0cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4913,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=68;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4913,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=simplai agent"],"subjectType":"system","unit":"percent","value":53.400000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-55acb18f0f005a5dc8ba","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Simplai Agent","harnessId":"Simplai Agent","httpStatus":200,"id":"pub-5be8b671057f2c161de40761","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Simplai Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.534","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fa5d8744a4803e1b977f3e903b4a93fd643b0cf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4912,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=67;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4912,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=simplai agent"],"subjectType":"system","unit":"percent","value":53.400000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d3db13ee25e58c799114","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-18ca020f217fb7a852e88e9c","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2026-02-10","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.396","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36938c47fc67b2fb2243049347a42abcecf73014","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4953,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=108;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4953,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":39.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-114cb587d5a274ca57f5","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-f330ce216d52360ddd29ee2c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16399999999999998","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140accf3b5cd9627acbf7750d9fe2599fad7bca6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5028,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=184;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5028,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":16.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c9c26fa9562e3d9a0fed","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-d05e3fb4919a267d5f163ea7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.171","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140accf3b5cd9627acbf7750d9fe2599fad7bca6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5022,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=178;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5022,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":17.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-62d260a7566fb7efcb4d","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-874ac16a60a169a4b3cb4b17","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16899999999999998","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140accf3b5cd9627acbf7750d9fe2599fad7bca6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5024,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=180;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5024,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":16.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b5281281056e0ea57234","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Gemini CLI","harnessId":"Gemini CLI","httpStatus":200,"id":"pub-4deaa0470368d110c5a00fe0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Gemini CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.154","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"140accf3b5cd9627acbf7750d9fe2599fad7bca6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5030,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=186;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5030,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":15.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa85b3ce6e8cd291aac5","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-cbd1853ec0a39305748f1e02","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.326","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4976,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=131;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4976,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":32.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-21a088aa56f17560c2d2","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-a62274d5c4b7213147f60525","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.261","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4998,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=154;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4998,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":26.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2435feabb8559c9ff394","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-9fa54e048636aae781460666","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.16399999999999998","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5027,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=183;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5027,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":16.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b9e1632128a9758ae7ad","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-9a5ccdccd1ea35e25c791153","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.261","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4997,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=153;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4997,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":26.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b73a89bf16eb3e271d59","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-6647c3ccc3ede3217381144b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.326","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4977,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=132;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4977,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":32.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d4f8eb8472206645858b","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Gemini CLI","harnessId":"Gemini CLI","httpStatus":200,"id":"pub-4d34337cdca371365319da01","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Gemini CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.196","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5018,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=174;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5018,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":19.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bf12bf385f6390226e40","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-4b0d97656771cb1b2dbb09c9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.164","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5025,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=181;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5025,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":16.400000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-13c450be821d69d8eec2","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Gemini CLI","harnessId":"Gemini CLI","httpStatus":200,"id":"pub-11172d4c06fca6f798839036","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Gemini CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.196","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"475f540abbfd46fee2a76e0543a6cecdfe0c8bbf","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5017,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=173;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5017,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":19.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-499a1b08a71ee06a3183","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus-KIRA","harnessId":"Terminus-KIRA","httpStatus":200,"id":"pub-a1298b1c5d967cc767587b0c","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus-KIRA","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.748","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"517487305dfc2bdf2ef44702f11c7e21b63b6438","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4865,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=19;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4865,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus-kira"],"subjectType":"system","unit":"percent","value":74.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-60fe91c524a5df3c3c0e","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"TongAgents","harnessId":"TongAgents","httpStatus":200,"id":"pub-92b82ae19a984604514efe22","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-03-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"TongAgents","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.802","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"517487305dfc2bdf2ef44702f11c7e21b63b6438","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4853,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=7;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4853,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=tongagents"],"subjectType":"system","unit":"percent","value":80.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a1262c3555eb5855e593","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"ForgeCode","harnessId":"ForgeCode","httpStatus":200,"id":"pub-5c766476f514b05d28686684","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-03-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"ForgeCode","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.784","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"517487305dfc2bdf2ef44702f11c7e21b63b6438","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4857,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=11;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4857,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":78.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-deada730b8bdaca79cee","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Forge Code","harnessId":"Forge Code","httpStatus":200,"id":"pub-3bb33446990f12e39cdd6552","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-03-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Forge Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.784","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"517487305dfc2bdf2ef44702f11c7e21b63b6438","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4856,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=10;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4856,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":78.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c727294ec7b517df7b46","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Gemini CLI","harnessId":"Gemini CLI","httpStatus":200,"id":"pub-06a50de460bb715d9c4ff562","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Gemini CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.594","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"517487305dfc2bdf2ef44702f11c7e21b63b6438","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4897,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=52;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4897,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":59.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32a47d54d8b5953e5288","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-71c2a9f3ea41723acab69fbe","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-01-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ceed8a7812a950f88b944184d3a9c866692119b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4974,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=129;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4974,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":33.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-24762e403f0f62aeac9e","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Crux","harnessId":"Crux","httpStatus":200,"id":"pub-4812140fa98deff916a16246","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-02-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Crux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.33299999999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ceed8a7812a950f88b944184d3a9c866692119b8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4975,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=130;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4975,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":33.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8663d9cf606f6692bd6a","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-a3aa5f42838ad454bbc38202","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.524","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9119754ef6bcfc9bc9c397aed63ba1ff63457300","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4914,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=69;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4914,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":52.400000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7e36f282807dc7e05909","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-f5790316f41b674f95505f1d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.339","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4972,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=127;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4972,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":33.900000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-709c4f8edcb14dd46d11","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-9ba75b776aba86feb0fc2476","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.352","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4965,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=120;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4965,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":35.199999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7657a3c1279b0c6e16d9","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-65af4d28d75eedf5253cf19d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.438","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4930,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=85;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4930,"metadata":{"epoch_file":"terminalbench_external.csv","notes":"Medium thinking used in other GPT-5 evaluations, assuming the same here.","organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":43.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3876d4dc369f0c461c45","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-031ea0ec0c11ffdaf834b509","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.496","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4921,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=76;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4921,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":49.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-05d5ef7a30ea3365824c","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-ed7444dcafd3720cd41bffcf","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.49600000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4920,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=75;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4920,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":49.60000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ef74ffb3674179ef85c6","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-b11dfb691b5e867fb9bd9e85","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.33899999999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4973,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=128;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4973,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":33.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-04e9ab079567d21fc163","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-a7b838da888b5a4a7e9c50e0","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.438","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4931,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=86;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4931,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":43.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":8,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ac2b1139aa90cee455f7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-7b88b7363ade5f727e3cd6fe","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.35200000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8205adc9fcaea688dce252e69097dab218e940","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4964,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=119;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4964,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":35.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-49fba6d6caf77bc51749","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-e141cd5ab208b7a12cf14c9d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.292","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4983,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=139;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4983,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":29.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32a2e43368b4bfdc485e","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-bd6e533c7f0b0726616b09a2","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5014,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=170;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5014,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":22.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-32954ccefa6f65889f6e","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-a4b76b1c3f76e64f185d0e6f","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.319","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4978,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=133;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4978,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":31.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3637a1253bdd3b5f68ab","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-481784c8c02fd50c86e00046","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5007,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=163;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5007,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75000d1cab3fdb4e178b","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-cebcd6e008ffbd619a1cf302","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5015,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=171;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5015,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":22.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-55d84e6b87d3ee13cc5d","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"spoox-m","harnessId":"spoox-m","httpStatus":200,"id":"pub-4cd604993013777c401f4402","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-12-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"spoox-m","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.348","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4970,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=125;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4970,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":34.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8be08bf0475c3d9921f4","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-462f5768a552a50d5ecf25f8","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.292","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4985,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=141;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4985,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":29.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8950a14114a5f0913f1f","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"spoox-o-m","harnessId":"spoox-o-m","httpStatus":200,"id":"pub-45f6f732cb647a26ed9be2e7","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-12-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"spoox-o-m","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.348","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4971,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=126;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4971,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":34.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-34e4713be3f15d2920e9","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-37126222e42a30618f649896","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.24","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5008,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=164;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5008,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c8131f3282924b663d71","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-2647775565f8c95b4a9a5cf1","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.319","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fa0a0baf2778345d6fbe3f8f605e60be3b1e362","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4979,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=134;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4979,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":31.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-499af6b1a0baab9457ea","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-a16ac277ea727da5466001da","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.079","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5042,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=198;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5042,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":7.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ee471730735a73e85b6c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-91bc8bc0ed283fbe8bd48991","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.115","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5037,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=193;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5037,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":11.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3844e2d138be65e919ec","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-3a5428aed69da8f431abd92d","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.099","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5039,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=195;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5039,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":9.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3ba3ce5ec9dd027d6bdf","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-0d85443c218e676cebdab475","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5044,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=200;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5044,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0485229ac7bb6e9523d0","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"spoox-o-m","harnessId":"spoox-o-m","httpStatus":200,"id":"pub-fe57af4e7a74dd2350318f14","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2026-05-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"spoox-o-m","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.218","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5016,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=172;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5016,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":21.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a38d9b13a0d6a6d820d5","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-fbaa4077ea0403a8c8f963f6","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2025-11-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.099","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5040,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=196;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5040,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":9.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2509c86d6e0b4656aaef","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mini-SWE-Agent","harnessId":"Mini-SWE-Agent","httpStatus":200,"id":"pub-e4c98b2f361c7b0a4603322f","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mini-SWE-Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5045,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=201;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5045,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":7.000000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6bad47af8a78d53b1f66","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-daba28d634432d8efcd831f5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2025-10-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.079","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5043,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=199;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5043,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":7.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":9,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-57e1da911488f536812d","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-c3718e7e69964d2376b2fcd9","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_unknown","observedAt":"2025-11-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.115","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e27a90a96a6d60eb6a8140f4d018d29d1f9429f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5038,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=194;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5038,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":11.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6f9d5ce6f46ed5734672","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-de3c6a0cd9f4b592c81949f2","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.476","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"103be879113a083b884bb8948ef84d72e89c3b2c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4923,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=78;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4923,"metadata":{"epoch_file":"terminalbench_external.csv","notes":"Assuming medium based on GPT-5 being run at medium.","organization":"OpenAI","release_date":"2025-11-13","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":47.599999999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9502e97dd583ed7314ce","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-21e90f53b8e6269b9df981ad","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.476","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"103be879113a083b884bb8948ef84d72e89c3b2c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4924,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=79;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4924,"metadata":{"epoch_file":"terminalbench_external.csv","notes":"Assuming medium based on GPT-5 being run at medium.","organization":"OpenAI","release_date":"2025-11-13","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":47.599999999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-afc3cea9fc4c8b470ce1","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-4b50a3ac57f019e7da67e308","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.47600000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"103be879113a083b884bb8948ef84d72e89c3b2c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4922,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=77;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4922,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":47.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8ea4d3c033fa50e45652","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Letta Code","harnessId":"Letta Code","httpStatus":200,"id":"pub-ff832f99704c74fcb6c5b5c9","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-12-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Letta Code","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdcd24e967ecda53f3f2ab42ecb879d088b12516","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4911,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=66;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4911,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":53.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-57850b34cdf6cb212dee","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Crux","harnessId":"Crux","httpStatus":200,"id":"pub-2e185f0c0231e98d7a6dc2d8","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-11-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Crux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.578","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdcd24e967ecda53f3f2ab42ecb879d088b12516","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4903,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=58;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4903,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":57.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ae3e517ab52228e747ca","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-223aaebafd05824e3c2ffcc2","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-11-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.578","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdcd24e967ecda53f3f2ab42ecb879d088b12516","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4902,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=57;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4902,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":57.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d138bd478a6843a2d361","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-1f669adfb416145c4296987b","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-11-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.369","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdcd24e967ecda53f3f2ab42ecb879d088b12516","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4959,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=114;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4959,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":36.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-accbc9dfa0183f9240c4","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-062cbe51489bf5c754891c03","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2025-11-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.369","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdcd24e967ecda53f3f2ab42ecb879d088b12516","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4956,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=111;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4956,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":36.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ac8014e6dcac193beda2","canonicalModelId":"openai/gpt-5.1-codex-max@2025-11-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-b67a257de5e068cb99b3d100","mappingCandidates":["openai/gpt-5.1-codex-max@2025-11-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Max release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Max release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex Max","modelRef":"gpt-5.1-codex-max","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.604","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6acfe8380f12ea744b3f91acd800721e89f735f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4896,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=51;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4896,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-19","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":60.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-289d130868d79ada93ec","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-9d135192f711eadafaa0d550","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2025-11-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.431","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed961b3e559fc9d87c842d015ff4ba31299c0c83","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4937,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=92;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4937,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":43.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-48628bd10b74580d81c9","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Crux","harnessId":"Crux","httpStatus":200,"id":"pub-9432acd6f93f494a891422a7","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2025-11-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Crux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43100000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed961b3e559fc9d87c842d015ff4ba31299c0c83","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4935,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=90;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4935,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":43.10000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7f6150e7d0b8ba4f1b91","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"hookele","harnessId":"hookele","httpStatus":200,"id":"pub-4c790dea209bc149bcc4122b","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"hookele","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.616","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed961b3e559fc9d87c842d015ff4ba31299c0c83","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4893,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=47;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4893,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":61.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-44e4ac0b31d9bb5a6bf0","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-384b5625c83ba60067a30e76","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.649","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72791d9941dd7db1451449b9bd8d190a5d461cd6","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4881,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=35;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4881,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":64.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ecd6082706afa3bec1a8","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-188516ff76d0cf439e9965dc","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.629","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72791d9941dd7db1451449b9bd8d190a5d461cd6","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4887,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=41;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4887,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":62.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-df028d2eaf8eb3ef304f","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-ac8604a0bd96988b3ede2789","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.629","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72791d9941dd7db1451449b9bd8d190a5d461cd6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4888,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=42;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4888,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":62.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-704f7c1195de30a0f4a7","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mux","harnessId":"Mux","httpStatus":200,"id":"pub-8681364fa8e94651ec469bad","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2026-01-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6070000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72791d9941dd7db1451449b9bd8d190a5d461cd6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4895,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=50;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4895,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":60.70000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-63885e6c4a67330897d5","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-7fd5dce78d388362ec6c86da","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.649","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72791d9941dd7db1451449b9bd8d190a5d461cd6","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4882,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=36;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4882,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":64.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":6,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2554dfe48691a30d65a4","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-2cb54dcab956c9d7a217d7e6","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.54","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72791d9941dd7db1451449b9bd8d190a5d461cd6","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4910,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=65;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4910,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2025-12-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7f9e6022c711df7f0599","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Deep Agents","harnessId":"Deep Agents","httpStatus":200,"id":"pub-3d0080f5d43551bed04eeaae","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Deep Agents","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"326ca06e0cdc952a512d4a9f2f3be3480d26327e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4878,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=32;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4878,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-18","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=deep agents"],"subjectType":"system","unit":"percent","value":66.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2650d281a58f901539dc","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Simple Codex","harnessId":"Simple Codex","httpStatus":200,"id":"pub-f4289c88f2d1992eb81c4ba5","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-06","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Simple Codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.751","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4864,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=18;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4864,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=simple codex"],"subjectType":"system","unit":"percent","value":75.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f7db9833d7a2ba3212e","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"spoox-o-m","harnessId":"spoox-o-m","httpStatus":200,"id":"pub-f1dcccf11e75d8fbd7e6bc62","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-05-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"spoox-o-m","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.715","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4870,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=24;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4870,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":71.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-08ea24a85f56f71f4e68","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codelia","harnessId":"Codelia","httpStatus":200,"id":"pub-e0d7e1522555df2a44b3bcc3","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codelia","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.757","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4862,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=16;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4862,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":75.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-544c5459377b48fb7aed","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"Droid","harnessId":"Droid","httpStatus":200,"id":"pub-df56a45d43539b9a1addadda","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Droid","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.773","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4859,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=13;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4859,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":77.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2b1fff68195f1aa8080b","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"CodeBrain-1","harnessId":"CodeBrain-1","httpStatus":200,"id":"pub-d639a48710811f2d9cf64641","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-10","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"CodeBrain-1","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.703","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4871,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=25;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4871,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":70.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b4c25e3fb0275140e35","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-c3a5662d7580a6c949716a97","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.647","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":6,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4883,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=37;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4883,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":64.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc283efd103423abf80c","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Mux","harnessId":"Mux","httpStatus":200,"id":"pub-baa7428664884ab66ddf959d","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-03-06","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Mux","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.746","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":7,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4867,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=21;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4867,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":74.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7527cb1e1115cffc0d3d","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"SageAgent","harnessId":"SageAgent","httpStatus":200,"id":"pub-b91597b34cf519e7af844991","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-03-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"SageAgent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.784","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":8,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4858,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=12;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4858,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=sageagent"],"subjectType":"system","unit":"percent","value":78.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2439c4a465b48a622f6c","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"CodeBrain-1.5","harnessId":"CodeBrain-1.5","httpStatus":200,"id":"pub-3d8eefb9abd3fcb3f76b683e","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-10","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"CodeBrain-1.5","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.758","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":9,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4861,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=15;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4861,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":75.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":10,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-46374bc4c7d1d73aeebc","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"IndusAGI Coding Agent","harnessId":"IndusAGI Coding Agent","httpStatus":200,"id":"pub-1d8508f36dbc42076b2eb714","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"IndusAGI Coding Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.691","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6a801b48c1b216edf9be96153a1a5533fde1be","selectionRank":10,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4875,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=29;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4875,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=indusagi coding agent"],"subjectType":"system","unit":"percent","value":69.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-02108d15267339eb208d","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"ForgeCode","harnessId":"ForgeCode","httpStatus":200,"id":"pub-0d9084acee734cd4bb5e8400","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"ForgeCode","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.818","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9aa667771a29df50cc79158d45dccec216d7ac6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4852,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=6;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4852,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":81.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-657b68b7e36d178bca77","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Capy","harnessId":"Capy","httpStatus":200,"id":"pub-e36d5144b0bab71e905a8e8a","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Capy","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.831","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06975f7fa49e430607e4de68eee8375d2f16cb63","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4849,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=3;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4849,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":83.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-732d82bf538fd53b72fc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex CLI","harnessId":"Codex CLI","httpStatus":200,"id":"pub-b760ca45b06556c138fd4cb1","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex CLI","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8220000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06975f7fa49e430607e4de68eee8375d2f16cb63","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4850,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=4;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4850,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex cli"],"subjectType":"system","unit":"percent","value":82.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-71380df95f2db50cbbcb","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"clnkr","harnessId":"clnkr","httpStatus":200,"id":"pub-655816fcab5ef07789cb4192","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"clnkr","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6609999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06975f7fa49e430607e4de68eee8375d2f16cb63","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4879,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=33;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4879,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":66.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6850c511d5a0d34e1fc4","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Codex","harnessId":"Codex","httpStatus":200,"id":"pub-55b7fae8661724a2f2001e9d","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Codex","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8200000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06975f7fa49e430607e4de68eee8375d2f16cb63","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4851,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=5;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4851,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=codex"],"subjectType":"system","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7983cccbb790d16d5f0b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"NexAU-AHE","harnessId":"NexAU-AHE","httpStatus":200,"id":"pub-0d475b1bec6ec1c307db522c","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"NexAU-AHE","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8470000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06975f7fa49e430607e4de68eee8375d2f16cb63","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4848,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=2;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4848,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":84.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c6ca773c24fc49e4e093","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"Terminus 2","harnessId":"Terminus 2","httpStatus":200,"id":"pub-2ec4f525feeb31b314baf29f","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-02-04","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Terminus 2","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.43200000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f0e0426e0c96c2ceff87a9145404a20731c6edd4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4934,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=89;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4934,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=terminus 2"],"subjectType":"system","unit":"percent","value":43.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b1a3863505b5fc630db5","canonicalModelId":"qwen/qwen3.5-9b@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"little-coder","harnessId":"little-coder","httpStatus":200,"id":"pub-5075b74fb1afdd316a71385b","mappingCandidates":["qwen/qwen3.5-9b@current"],"mappingEvidence":{"confidence":"high","note":"Epoch exact capitalization for Qwen3.5 9B.","registry":"model_aliases.json"},"mappingNote":"Epoch exact capitalization for Qwen3.5 9B.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3.5 9B","modelRef":"qwen3.5-9B","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"little-coder","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.092","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36ad1961f633a19ca408805ec3a5966381787cee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5041,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=197;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5041,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":9.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-terminalbench_external","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a6f90aabad8cc0788760","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":"little-coder","harnessId":"little-coder","httpStatus":200,"id":"pub-5f298746fb7cb46820d5df3b","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy mean","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-05-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"little-coder","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c63c3820b0c27fbad783717d95b56933ece4b0a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5013,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"terminalbench_external.csv:row=169;column=Accuracy mean","sourceMetricId":"Accuracy mean","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5013,"metadata":{"epoch_file":"terminalbench_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":""},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":23.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-the_agent_company_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0752448da771d33fa6c8","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c3f24d5c5cf47ad8a91385d","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"% Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp","observedAt":"2025-10-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.524","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1b67e5d9c43aac3d2a715fcd292061e7c4b8794f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5061,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"the_agent_company_external.csv:row=13;column=% Score","sourceMetricId":"% Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5061,"metadata":{"epoch_file":"the_agent_company_external.csv","notes":"Not confident on organization name.","organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":52.400000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1b4bf113460dc1ac0546","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-48c103db4283bf7e651a77c1","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1034.0040000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10e13193d62635c96dfc2845055df8c0f35d6a45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5225,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=46;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5225,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1034.0040000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4f0b90bbfc84c66d8283","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-df00b9c3a14d094e59929852","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"-23.15999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f38a94a7e2a33693b13b426fefd132868e775d7c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5238,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=59;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5238,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":-23.15999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3bebd361a27d39ec5df6","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3dc6db12d3ed61f0b39b33c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2157.7716666666665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d995f697494ddbb457fd6679b93b799a4eebcb0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5221,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=42;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5221,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2157.7716666666665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9e93910384fa8ec77bda","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c69dcb8f4d572bed9d7544ed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4529.944000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"414b0e03c6e13825734ff0c041c4b9292dddee67","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5208,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=29;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5208,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4529.944000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-abd809319763a37961f5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-15c18a98b076327348ac78ec","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5680.258","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"414b0e03c6e13825734ff0c041c4b9292dddee67","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5197,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=18;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5197,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5680.258,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d7bee6f3939b045e69a7","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d76ac35cd74738b40462b8a1","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5018.519999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"414b0e03c6e13825734ff0c041c4b9292dddee67","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5204,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=25;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5204,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5018.519999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4184ab32eeceb0a07059","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c64b09e9686886c6dd4b9b74","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4966.639999999998","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"414b0e03c6e13825734ff0c041c4b9292dddee67","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5206,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=27;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5206,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4966.639999999998,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-07a47cb8b2bc1890d450","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fea4823425165dc753642c6d","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_medium","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4339.812500000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"414b0e03c6e13825734ff0c041c4b9292dddee67","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5210,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=31;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5210,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4339.812500000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-df09fa7e1647b9371d32","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a45d84b0b13b73e3d0dc117f","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"458.88599999999985","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1090aaf85f9fd9bf8c8565ba3acceadda985f1cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5230,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=51;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5230,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":458.88599999999985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca97c0cb3841b707758e","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5c0952d20fa2b6e95c8d60c9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4967.056666666668","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5cbe002a6224807874df8a11166e8bf9e1a4b792","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5205,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=26;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5205,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4967.056666666668,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0e270e3a9e7b975b8dca","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f2b48f7aa60d88993523517f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"8017.594","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df655febf92ec33d7ff7c04dac2d90641e642d80","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5187,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=8;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5187,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":8017.594,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7710d44e72db7ed93834","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-63d714cbb90889eb8236afc6","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"10936.763333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4428a7230613a890d0920a227647ae8e33f33d48","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5182,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5182,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":10936.763333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2bb5a75efb75f11a8b55","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-413ce34198252a034be8a2b7","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2992.344","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"642d0a23534dc487e6b582eaa0993153a51da500","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5219,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=40;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5219,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2992.344,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-94df10f0714180980e8e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-911712ecdb3689b33266c532","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5787.428","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"642d0a23534dc487e6b582eaa0993153a51da500","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5196,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=17;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5196,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5787.428,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5d5dc1b60a795a20309a","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1fefd93818186c330feb63b","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 5.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_unknown","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"11181.87333333333","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"893026b7aa8372af5fc27cac41160c672a94b810","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5181,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=2;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5181,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":11181.87333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-97d311c75cb9d7eb5961","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-07dddd746fd44147f0a895bc","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"3838.7400000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c9d603d8153ff70bb355f048e17287f3326a468","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5214,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=35;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5214,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":3838.7400000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-46d6d4f4cbfe74619095","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-82e6ec121c2ea45b6cc39c1c","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_unknown","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"7204.139999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58e652bb59d23340ef3c50d7c96d0e6d5495041a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5190,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=11;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5190,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":7204.139999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7b9fdcb935400c7e3c12","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c764cd8fe4dd09783c5f1449","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 5 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 5 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_unknown","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"6377.701666666665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b5708d3c455cb7f2c81dfd59bcd0d525de2354d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5192,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=13;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5192,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":6377.701666666665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59c4e68706466ab928c6","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6db72572f0637803753a965b","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_unknown","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"3284.519999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b82b247176305f5de9d941db23efe24bff075af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5218,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=39;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5218,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":3284.519999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-218295ae4e61d0b4b170","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-afe0d9081c13e2fb7e2f57c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"548.836","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21b6cb4e923cee37e1939d7a2cd2dfb4e216f60e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5228,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=49;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5228,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":548.836,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0bf022c02aa2769d368f","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5d3efbe88361aaca994051ea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"573.6420000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3cd34f210769dfb0dc10c48fa9f1352dbbcf439b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5227,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=48;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5227,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":573.6420000000003,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-dcd669d32dfd67a21806","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f534c925bfad82f154570182","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"911.2139999999997","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1351a57c500c2aaa1582bb4dfd00d7936187877d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5226,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=47;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5226,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":911.2139999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8de6aa2e5a39a800e993","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-347378eb0350356905112977","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5396.418333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92155aeed08ab64db77d0bf54953f958b16bb62b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5200,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=21;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5200,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5396.418333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c13f5b7a9b8b382eb73a","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2137ebc007e8e973606b47f","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"2376.816","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b8d1088bc9703e7b5e1ca1f84ed8363dd2fa7467","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5220,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=41;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5220,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":2376.816,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a6e9ed6d9af31f53b550","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5e3462b84705f62ed1eb8c8c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4432.122000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2f8c3a93d999c59dd86ef4088d5a85fbfd66b7c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5209,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=30;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5209,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4432.122000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ab1181681be8102c6621","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1911749fb90020c58766e95d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5634.406","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"467e41241d7cd6325a8d2780dcf01b8b6d9f5d1e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5198,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=19;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5198,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5634.406,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e982946c00bc5894932d","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e733952299d5e89dc23b9a4d","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.2","modelRef":"glm-5.2_unknown","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"8313.778333333335","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"287e4602ad5d39bade87e9fd5bd402a607378727","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5185,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5185,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":8313.778333333335,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9acfd3447de8820ff7e5","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7b7534495c6de64032959244","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GLM-5.3","modelRef":"glm-5.3_unknown","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"8163.609999999998","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a4661c5f03b28e06d6359fc5ca3dc86820a6200","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5186,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=7;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5186,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":8163.609999999998,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad2eff0a134ab08a5aae","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a1e28f66fd9963d38d4fc424","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"-31.18399999999995","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3ef3ab3a9d0c78a7c6ae1fdea6995010b97cb09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5239,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=60;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5239,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":-31.18399999999995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c421f2ecfd92836b0301","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8a8324610e0b7576cc6126cc","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1473.432","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41a7c94d60acc3ee5f1557a1e0612898a4638702","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5222,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=43;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5222,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1473.432,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b1ffac1a08e1ab0fe0d9","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9f7fe6a7da6e9ad26c2d1e7a","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"3591.3259999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0bd24ce7684cfb27b91a0993eb8f2e171d4ef3cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5217,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=38;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5217,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":3591.3259999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2f231379019aac388394","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-43eb37a8c07a4f8366869a0d","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5940.124000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42fea67ea3462931924265fcd8e93af0447800ae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5195,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=16;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5195,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5940.124000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-76bb47226e64e2919916","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-63893b9efe1bfacbeb418364","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"6144.176","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34a69e194f91f51da67b3db0307056d8e82a5328","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5194,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=15;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5194,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":6144.176,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-28ce4d7a24326a324d30","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-648b2b4e6f4cb5797222dc3a","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"7523.843999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5750125f943c9acc27bd6a45a6463cf15f90e414","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5188,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=9;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5188,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":7523.843999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4c0b453f1ac0396f8f5c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-21a4eb23142f1d6041cf9c29","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4094.7120000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"615b310177bb050473cb413b3a80001488cbe083","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5212,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=33;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5212,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4094.7120000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7065341af6c6c001f02e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cdcfa46a82adc41853d2f906","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"9619.367999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04e8a9279801583dd7d08fa092dd392a86cff24d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5183,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=4;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5183,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":9619.367999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-76ad3284cb6c1cd61a50","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ce6523ac8ad811bd9f4465e1","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_unknown","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"7343.206","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a75bc6c9ea203323365bcde7644a9e8b4245c4d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5189,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=10;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5189,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":7343.206,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b4fa5e03e4db14572a8b","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1e46ddf23e0ea3d4b389d862","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"3887.4259999999986","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"062748cb37486fef79780d60c0666907716d4ce2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5213,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=34;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5213,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":3887.4259999999986,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7215d5844d96fb3c0693","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-16ba2d73b562756428d521df","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Grok 4.6","modelRef":"grok-4.6_unknown","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"9047.032500000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8fd1dbcd9cb7ec3b2fcca77b3a4680af7cd3eec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5184,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5184,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":9047.032500000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4505750dbc5967d414c3","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b54e809f3ea3abc56f17cf1","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"1198.4579999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7411f09a5893bcf5f9addc3caa24df7876b5f38","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5223,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=44;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5223,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":1198.4579999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-03ce719f22fcede0d7c2","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88eb81cb96d83a12475f664d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"6204.570000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9a159e070409ebfbfbd156902d9dabc8161c261","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5193,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=14;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5193,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":6204.570000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c37aefe9d99db2b879a8","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bfab718ca974d8c592941def","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5082.936","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e151a90d411f4740b33dc85e7bf9bfb53669324","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5203,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=24;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5203,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5082.936,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d7cc91c1c9e426df0bf","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0912338dfab9b9cb0075765b","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Kimi K3.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Kimi K3.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K3","modelRef":"kimi-k3_unknown","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5165.035000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9911dc09f23aa7ce1622eb95649bdd9c0bf51fef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5201,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=22;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5201,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5165.035000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f257edc6695eb16af539","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-25170f3786f1b93186e48d2e","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27B","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"201.97599999999994","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"892f552c6e38e89cad85a9a6ad9320870984b6d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5231,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=52;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5231,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":201.97599999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b626ae11838f3aa7c9bb","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f14d2ec0b0780dbeeecac235","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"462.68500000000023","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a6cc71702b12a70711f3db2ea4d9ec6cb062898a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5229,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=50;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5229,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":462.68500000000023,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fe265bcd5a729c5e2d5e","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0d0fdbdd7eb4ae0accb40ff5","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"0.5440000000000417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90af44cf0340ff78b5b762de8c8f75dffa46ee50","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5235,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=56;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5235,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":0.5440000000000417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ec9a6c60dbca55e73f60","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-794b8fd3cedf6803b7b17970","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"4254.1866666666665","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"690a74c35220112170bdc18c54129def5ded2dd4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5211,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=32;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5211,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":4254.1866666666665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vending_bench_2_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5805b0519455a10da303","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2027dee135d9ed0ef608baf2","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_unverified"],"rank":null,"rawValue":"5114.872000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8188fb23b3feaa17a1b43b2a4aafd3e9a73f86db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5202,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vending_bench_2_external.csv:row=23;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5202,"metadata":{"epoch_file":"vending_bench_2_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"score","value":5114.872000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b52c7ddf32b03dbcd1ff","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-27e68d495d256d085102769c","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.33","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c3363bab0fa4823ea747a065479ad8a2b1628df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5273,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=35;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5273,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-26168f3cfd59784e0232","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-00ec2589ea0a236cacf8d8ad","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c3363bab0fa4823ea747a065479ad8a2b1628df","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5263,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=25;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5263,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-743116b3628b1de17da5","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://x.com/ChaseBrowe32432/status/1993038782975951149","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-50c6439828295849dfad3d8b","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1617d273b53749dd345417eb8e84ee6023736145","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5258,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=20;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5258,"metadata":{"epoch_file":"vpct_external.csv","notes":"Hasn't been added to leaderboard, but twitter thread confirms it.","organization":"Anthropic","release_date":"2025-11-24","source_link":"https://x.com/ChaseBrowe32432/status/1993038782975951149"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2b4345418b562fa61fc9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c4a505b520f2b03c6160ba8","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bf851e3413906508ecd96818cdadffe00175f08","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5264,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=26;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5264,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b0c0959efca358f93e9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-22e68fc2435c44c601473252","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.398","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bf851e3413906508ecd96818cdadffe00175f08","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5259,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=21;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5259,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.800000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d1ffe0d642e431fafa03","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cde7d01f05852e28356e780c","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.66","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9fc504d35ee935409c63a69adb83124f6636a471","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5244,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=6;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5244,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6d65972ce85aab0f60cd","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0424cae00f95fe7e07a345af","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.632","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9fc504d35ee935409c63a69adb83124f6636a471","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5245,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=7;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5245,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":63.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3fd278fb87f841bbb7b1","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bc7d22f4916868c21eb86d2a","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.39","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35b69831777a836ab9e110bba532912696f25a8c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5261,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=23;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5261,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e24c58de1c59eb661b99","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ecd66aa0ba805a074db62a92","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.402","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35b69831777a836ab9e110bba532912696f25a8c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5256,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=18;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5256,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c5ce19ba5e4f23fa0744","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ef4449402efebe1839d827ee","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.372","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ade6350a3b24b35ec9ec5e0cff7ee9c1fa1cb4a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5265,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=27;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5265,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9603345f8025fbf6b5ab","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3d6d7de09e08db84872e76ab","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.354","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ade6350a3b24b35ec9ec5e0cff7ee9c1fa1cb4a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5267,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=29;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5267,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-16f78cf6964ac8961c6e","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0288541cd8ba1db33aae5805","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.587","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce75d4dbe18e0246ab94740f25deceb43a793fd6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5246,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=8;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5246,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4f9dd4cac991c7badbae","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-46506a86b6564996b98a141a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.533","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce75d4dbe18e0246ab94740f25deceb43a793fd6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5248,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=10;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5248,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.300000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75d603f1e981437f7c79","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5856285de606a2196803dc20","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.67","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a010ef5c37b0f4a2baa4e0071367d296de2f9de","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5243,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=5;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5243,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":67.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-vpct_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c72685e13af9669fcf20","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://cbrower.dev/vpct","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f2d022b1b6b3a187950f6781","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Correct","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.84","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a010ef5c37b0f4a2baa4e0071367d296de2f9de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5241,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"vpct_external.csv:row=3;column=Correct","sourceMetricId":"Correct","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5241,"metadata":{"epoch_file":"vpct_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://cbrower.dev/vpct"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c00e513bb0f3220f5bb4","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f3576e4a85792ef63a5f6ca","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1271.77","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"820190ffcf862cad3f9ec367ea24eed1d95146f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5381,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=107;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5381,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1271.77,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4d0efe0065d080e89e63","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64b18a44cc1c7c8a411a078a","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1367.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"820190ffcf862cad3f9ec367ea24eed1d95146f9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5361,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=86;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5361,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1367.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-125aab233a502eb55d38","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e22bc38c5cd245ea1edf89ac","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiniMax M2.1","modelRef":"MiniMax-M2.1","observedAt":"2025-12-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1386.99","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f26e5d58ffa22a419539bdea9c54f63c65b6a3d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5354,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=78;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5354,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"MiniMax","release_date":"2025-12-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1386.99,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-cfba650d23b7b54d3e00","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88233b798cf65ebd025bb4f3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiniMax M2.5","modelRef":"MiniMax-M2.5","observedAt":"2026-02-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1383.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"294f494244453435803e34484f751b86e90f199c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5357,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=81;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5357,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-02-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1383.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bcbeed9751af7ba354e1","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ce2316dc3024bd677134747b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1398.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c81a7792c3dd642f5c0b0463ca12434fa2faddae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5347,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=71;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5347,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1398.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-06-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2dfaebb6c5ad8b8b2ab8","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-55bb03ba7ffc1ac53c613e56","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1487.64","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ab817947d69669434b8efc349ee792f0fb0b2e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5318,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=42;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5318,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1487.64,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0ac35488c8f1501ef98a","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ab8ed62d0405eb95de75ab2","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1273.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5de3264051546efb89bbefdc9809a7943fd1bdb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5380,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=106;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5380,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1273.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-381a2a0db44a1fc8afeb","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-66093b507d7e2457817d5fd7","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; Fable remains a distinct restricted catalog release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_unknown","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1626.26","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"705cf08b5332117b15dd01e62a0b93fcdff066d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5283,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=7;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5283,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1626.26,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-303bdb8c9cd4926b78a5","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4a6e6ae129ec78fe17aa6b10","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1286.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"39e15aa2a3042fc2b95eac6a24ddcd5951ea64a0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5379,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=105;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5379,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1286.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5b45aea6475af1789095","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53fbf1742bc08dbf9aa5b66f","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_unknown","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1327.41","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"39e15aa2a3042fc2b95eac6a24ddcd5951ea64a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5375,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=100;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5375,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1327.41,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2a3daa97159f0d54cb80","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2a0e9171a5886b01a8a53ebd","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1468.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"118acc17203416722a9d19704d487c4dfb637e4a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5324,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=48;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5324,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1468.08,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-061f0c5348e4e949fcc9","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-86e76c854e9a7e7ca0ecaaff","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1494.39","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"118acc17203416722a9d19704d487c4dfb637e4a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5316,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=40;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5316,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1494.39,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-498d9bb9967df4bddad9","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80cd5c8041888031d2ee35cb","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1536.27","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7213ad574ee80912763a90ad4f31f3e25b54cb70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5303,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=27;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5303,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1536.27,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ca68315098d1924904de","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e9be1a187d0ff284da1d22e","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1545.68","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7213ad574ee80912763a90ad4f31f3e25b54cb70","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5298,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=22;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5298,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1545.68,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1ef8f1976476ad9ace68","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2db44023330516536d5864f9","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1556.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7213ad574ee80912763a90ad4f31f3e25b54cb70","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5296,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=20;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5296,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1556.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9eaf3ee8bfd59f0d2fd2","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-01dc81c524664e3637f41fdd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1557.52","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69e9e011c12035b4e747043ce1da766146fff654","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5293,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=17;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5293,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1557.52,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b3ae84a855bf95ad31b7","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fb6b122af31d5cd417be90f0","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1556.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69e9e011c12035b4e747043ce1da766146fff654","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5294,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=18;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5294,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1556.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6508dc53c46abb4d0e56","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bd02722db2912d099fb1477a","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1566.85","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69e9e011c12035b4e747043ce1da766146fff654","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5291,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=15;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5291,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1566.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-50c8ac5083e8aae2ed49","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6976388f8758a75f5738658","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_high","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1563.49","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4b0cd3299ed6a5b4e2ba88aa58a6460fa8bf90d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5292,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=16;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5292,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1563.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-731e6ab2bbd92cf8f31a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f00141e8e8d1516b8ccd7b4","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1539.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4b0cd3299ed6a5b4e2ba88aa58a6460fa8bf90d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5299,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=23;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5299,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1539.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a8d230282de72bd1ecc4","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c826f3330d184f45995679a7","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_unknown","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1552.22","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4b0cd3299ed6a5b4e2ba88aa58a6460fa8bf90d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5297,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=21;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5297,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1552.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d01d5c79df8e731c4380","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77a546a442d6f55538df6971","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1662.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6275a30ee983afea03e79ae5014146703489a5f2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5281,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=5;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5281,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1662.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ad81e321547e439e2c85","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-35932c1b56ae9221e273bac9","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1690.64","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6275a30ee983afea03e79ae5014146703489a5f2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5278,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=2;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5278,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1690.64,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-18e2af5ec36491165fdd","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-160525ab4acd0f1458dfab0c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1385.51","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d81f77729dffd471063c130f919acd4c46aba12","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5356,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=80;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5356,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1385.51,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-572b9f235bd52d892287","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b0a5137e9b28993a463a4170","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1391.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d81f77729dffd471063c130f919acd4c46aba12","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5351,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=75;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5351,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1391.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-30f7b3ce4ef50f538a60","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bbdd67ca6c3e54c825f0715","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1353.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d81f77729dffd471063c130f919acd4c46aba12","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5367,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=92;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5367,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1353.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-75198b655a24de7ebce0","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f044e22b554f42dcbf36fbb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1522.47","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f06cdbeb7b2768cacbda7d86e87e2d92f21930f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5306,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=30;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5306,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1522.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b19575cfcf7446b898a3","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-684a7a955a0a86d8410a7212","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1538.58","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d0d8542283c61167f00c538814e03f672ea43e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5302,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=26;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5302,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1538.58,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c298e3e69a578a7b1366","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9abf145fcd16e3e380512a14","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_high","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1579.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd0a1e22a64d4c732be972e82d73f0c3216443fc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5290,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=14;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5290,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1579.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-15c82ad44f0fcd9e4af7","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9b32a80cfafbbdb96659349","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1431.07","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd0a1e22a64d4c732be972e82d73f0c3216443fc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5340,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=64;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5340,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1431.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-61374f6da8c756776701","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88a1b0e4470e8f6779f2a96a","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1581.64","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ce5102e5af62c00d59a580c0305b750d837b620","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5289,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=13;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5289,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1581.64,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-16778013629869ac3388","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a208f2c971d26c9194f03f62","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1463.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5572c1177316d982a995741d41ba7f38ba544717","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5325,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=49;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5325,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1463.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-71e1b9c805b5e29a81f5","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe244b2105db24f47c8f2e1b","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1445.46","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5572c1177316d982a995741d41ba7f38ba544717","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5332,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=56;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5332,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1445.46,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ab87ca6b5f7af267ecf0","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-abf5fa4eea2107aa7158633f","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_unknown","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1463.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5572c1177316d982a995741d41ba7f38ba544717","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5326,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=50;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5326,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1463.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-01","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a49f72636eb4d303db60","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fa78cc7d41199e52174f9020","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"DeepSeek V3.2","modelRef":"deepseek/deepseek-v3.2","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1324.54","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"826e7d793f2a59161f221f558a670092febea169","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5376,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=101;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5376,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1324.54,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-06-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-be126b5ee13789dc3a00","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e96dabf103a2ee1541103fb5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1225.69","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a799eeee18be0fc889374b5b6fd65df66d78119","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5389,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=117;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5389,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1225.69,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-87f1ccb355c1e821fe95","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-790c8e208632eb20b088bec1","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1446.46","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0c21176d4c66d4b065ce23ed38844240c80fd57","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5331,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=55;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5331,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1446.46,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-597663e65ea75885d60a","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-83ea72ef77f83e0454942c86","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1449.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc0ffc7f6433e7a62e35ccfb1e2c4f57b83f3270","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5330,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=54;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5330,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1449.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4afe145f9a4a792d6337","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-85c95df727d7549c8783580b","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1498.59","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50be2e5b271ef90548b65761e97386f262748c52","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5315,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=39;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5315,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1498.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4811710dbe4c0d1e3e30","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1e80f26300163cf04e342be","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_medium","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1489.74","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50be2e5b271ef90548b65761e97386f262748c52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5317,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=41;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5317,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1489.74,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0b8714aa170d3916ad3f","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7583829ac8634940ebc561d4","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_unknown","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1506.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50be2e5b271ef90548b65761e97386f262748c52","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5314,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=38;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5314,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1506.38,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6562ed432d6506982fe0","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-235fe21b00748fb4dcfad298","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1538.91","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d14ccba8239be445e6a067a45d4e3b346ad5f6ff","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5301,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=25;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5301,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1538.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d6152cfc0506509b1d16","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fc49cfbc1b1589a8dcaa8a94","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_unknown","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1527.83","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d14ccba8239be445e6a067a45d4e3b346ad5f6ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5305,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=29;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5305,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1527.83,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-41a03f4f1722438d3145","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c5b56475faea4e56f3c77e6","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1587.04","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7526c9fb5cafc4e2f551f45f519005c73be8869","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5288,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=12;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5288,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1587.04,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a07f493c2f0c66c4605c","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2e3e8b3f6d615dc5bff47f5c","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1434.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"995f80a3fd5982e87c95124a7e7557693868976f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5338,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=62;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5338,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1434.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-227095b5970870ad127a","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a8d1ed720d6ec24ebd66e122","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1435.46","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66528ea8168cfb4e61f2c87ceb46dafeac83eaf9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5337,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=61;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5337,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1435.46,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-558190c63c0d1065a3d6","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0583c049080a1e2e404c59b6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1508.93","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"015c69adaa4ab2c19153d6aaf75b130daf0d68a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5311,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=35;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5311,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1508.93,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-aa794c3e9b78578c3f75","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9ff2cce9a9e198157e612975","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1593.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d5a54ce90198378f7bdf231133d6b1c66f0cf9ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5287,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=11;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5287,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1593.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9bdcf454fc236c9d4733","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bcb9a5cdaaf2cdcdcabee4f8","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GLM-5.3","modelRef":"glm-5.3_max","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1598.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0ca4dd1d8852c95f2b90f7ad35112865fb2494e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5285,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=9;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5285,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1598.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2b810a4e33dfe34d595","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fa5c7d5da3a102e05b064345","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1419.14","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d08c8eb5b3f334810fedc5c002e48940561c196","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5341,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=65;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5341,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1419.14,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f2712f3a6dabc33fba96","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7c954529b472de07ec2b3485","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1359.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2206f0a919413c079b4863f3eae062b077c22d2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5364,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=89;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5364,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1359.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-00de25dac2ec545e030a","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-472c1914c29f547ceb40d898","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1391.01","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2206f0a919413c079b4863f3eae062b077c22d2","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5350,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=74;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5350,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1391.01,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d3071350a76d32632012","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec0096d5dcb6e7c8c241210e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1340.63","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2206f0a919413c079b4863f3eae062b077c22d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5369,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=94;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5369,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1340.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-81f3d8ca1c4920fdc4fc","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-de5641f0fba652254f991676","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.1 Codex","modelRef":"gpt-5.1-codex","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1336.28","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36df1dab1d827da1450a20251fd810ce848c1b75","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5373,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=98;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5373,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1336.28,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-11-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1761b640db0756987310","canonicalModelId":"openai/gpt-5.1-codex-mini@2025-11-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-208e79e7dea20ce65cf7de0e","mappingCandidates":["openai/gpt-5.1-codex-mini@2025-11-17"],"mappingEvidence":{"confidence":"high","note":"Explicit Codex Mini release label.","registry":"model_aliases.json"},"mappingNote":"Explicit Codex Mini release label.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.1 Codex Mini","modelRef":"gpt-5.1-codex-mini","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1247.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3dc208ae03167f39df9c4efbc64ba4bcf91c824","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5385,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=112;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5385,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1247.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b5a464e5a8d1a572fd3a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7a6b384dd2c50015296582e2","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1480.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40751a5fc1cd64c401221341e2e9d9cf8fec3541","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5320,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=44;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5320,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1480.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3f4146ad7597d471bc71","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3d2874c9cb26e93d7968fcd","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1394.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40751a5fc1cd64c401221341e2e9d9cf8fec3541","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5349,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=73;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5349,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1394.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ef22fbe58907e5f0aad7","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2482631cc1eb18cdd64d827f","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1417.81","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40751a5fc1cd64c401221341e2e9d9cf8fec3541","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5342,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=66;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5342,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1417.81,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-12-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1edc7ce06c9cb826a802","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-771e474c3e78b0de496215ae","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2025-12-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1338.28","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4e1e5e20f4b1981d55340750ce60e73fca110deb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5371,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=96;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5371,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1338.28,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8acf8371c7314dd887ba","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d43a1fde529a07e948835c4b","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1369.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f60c1466926a62c4ebc81eb5d4275c25a30a967","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5360,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=85;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5360,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1369.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1735d56c36fa7b987870","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac350eff854823115f1a9e4a","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1462.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2949844586d0db23ce464597127dd84c74e8822d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5327,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=51;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5327,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1462.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-81073ed7cc15c5ba88a1","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e6976acafa1e7d0061a4b0b1","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1442.16","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2949844586d0db23ce464597127dd84c74e8822d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5333,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=57;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5333,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1442.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c836ca7b21c6132686fd","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0c03b913abe5a362d9f44935","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1390.56","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2949844586d0db23ce464597127dd84c74e8822d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5352,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=76;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5352,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1390.56,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-10d7d460f039a3ce8c44","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-811a0cbf7ac324f979ac8d49","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1397.08","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad24bb7cd4cd20a7177a2398f6e1e6f60ec2bc1e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5348,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=72;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5348,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1397.08,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c72b724ee6d97e896ddf","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-105df3847889560843ef96f6","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1486.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83caaddec25bfc1916f04c1237811db182bb7b62","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5319,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=43;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5319,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1486.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3eb29563e9ca9ee3f5dd","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-575bb82b8abd5aa7d5bfef4e","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_unknown","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1457.51","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83caaddec25bfc1916f04c1237811db182bb7b62","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5329,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=53;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5329,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1457.51,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-12839173c824a8e46426","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2c32ebf4d3b56221f1623b9a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1508.35","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83caaddec25bfc1916f04c1237811db182bb7b62","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5313,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=37;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5313,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1508.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-303e119b8b1c17ff826c","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7759c88c0476bb50fcf877e4","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1517.6","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aad516928008539868da6835bd35dccddb16e82b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5308,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=32;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5308,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1517.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b13c0330292ae512a420","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3813200b239a7e973640cf87","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1618.71","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84f9733b6cc38c46eb16d9c9adccf0ca29aaa181","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5284,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=8;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5284,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1618.71,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-acfc4187acdb66737989","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e0502b213415f5492c56fe8f","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_xhigh","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1519.87","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b23dae2a0f16c23dd237234b48c79d5bb2fe1c6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5307,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=31;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5307,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1519.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2025-09-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-afe2fbdb849efd79a848","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2609d18b4db1bf8b2382620d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Grok 4 Fast","modelRef":"grok-4-fast","observedAt":"2026-01-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1149.0","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0575113c5d3fb166989a95c1e146c79ed95f6d0b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5394,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=123;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5394,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"xAI","release_date":"2025-09-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1149.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-59afd36d0f4a2f4f1c4a","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dbefc6281965d6d316ab145a","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1556.36","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"763472f135bbebc191bc09ac22067e2d20f75865","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5295,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=19;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5295,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1556.36,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-51651303bd0a1484e3aa","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-88cb4c008ca1ae3b1c521757","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1628.84","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07020fb94a57acad0b35294a9654e8f7cec77f88","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5282,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=6;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5282,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1628.84,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9631e738a3bac769ec3d","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c652cfeb6cc920c7f9a2a61b","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1435.92","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32dd27ea655bc031b0571bdccb6d29c2df9af25c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5336,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=60;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5336,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1435.92,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-51f4236a0516e13de1f9","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6136a717a0080bd960e75a23","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5_none","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1404.86","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32dd27ea655bc031b0571bdccb6d29c2df9af25c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5344,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=68;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5344,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1404.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f0078ec22876bccf459f","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dba80fd728c27f4ce2fa4351","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1508.71","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f886052dbb7ae22e0336faebbfa117d8e999a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5312,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=36;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5312,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1508.71,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5f01347234077d8ce1a8","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5c96ff3346d130abd1e5bc64","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1472.65","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47c2c873a553f4c8aff2a51a8e6c696602027771","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5323,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=47;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5323,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1472.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-92b402849994a0ab37f9","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-54040bf5cc5f335e9f6d509e","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1673.7","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a83606141361045dfef2f3f3aad1996b22a64c23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5279,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=3;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5279,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1673.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-b0b242dc624ed902160a","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30b7f1c70d2b51e53d2ef553","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiMo V2 Flash","modelRef":"mimo-v2-flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_inferred_from_column"],"rank":null,"rawValue":"1330.17","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3103b4a54eca119c37d0414b429bcd2da005ad6f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5374,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=99;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5374,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1330.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-9ebf459137916206f02d","canonicalModelId":"xiaomi/mimo-v2-pro@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-662649f11b6a639f490e5e2a","mappingCandidates":["xiaomi/mimo-v2-pro@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiMo V2 Pro","modelRef":"mimo-v2-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","unit_inferred_from_column"],"rank":null,"rawValue":"1433.51","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fe257a67f5ef7f6205c7c396c3d5a8d12277a18b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5339,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=63;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5339,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1433.51,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-cd687955252dc73a8ff3","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1e727b380cef7aaab1e87b4a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiMo-V2.5","modelRef":"mimo-v2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["unit_inferred_from_column"],"rank":null,"rawValue":"1437.8","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7ca8e3730f834964af2b618b86148df952b9acd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5335,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=59;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5335,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1437.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-src-epoch-benchmark-hub-0c40568255cfdefe6926","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-289943ecab5faa7729b6b8fe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"MiMo-V2.5-Pro","modelRef":"mimo-v2.5-pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["unit_inferred_from_column"],"rank":null,"rawValue":"1475.65","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bac44a7be9f3ec90523e0661614eab40a3c78622","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5322,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=46;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5322,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Xiaomi Corp","release_date":"","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1475.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-239204a56797164b1dee","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6416c4bfaa89ec6ed12805b8","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27B","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1357.46","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c1da131246361238c380a51084ba6f4da77aac4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5365,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=90;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5365,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1357.46,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-985f3d9ee935a6032625","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4616cb178134e0757133c488","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1238.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ff4000a7266564f0bef3b8535a91bcb3444f655f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5387,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=114;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5387,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1238.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-02-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bbe0a2d93b2b622ddb7a","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac1555a08ca310a7608f8cc9","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1399.23","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf115d85998fe81a57d7267ad3ed2bf661bfcf69","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5346,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=70;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5346,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1399.23,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f762f1c98b20937e04ac","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-56abe9e2e9fb4145bf3fba03","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1249.75","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52655442b516da67a67da0c76dad6e5ed053673e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5384,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=111;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5384,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1249.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2d2c3b00daa52d038785","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-673da282ad4290da08da3986","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1478.88","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"def81301038c9f562e5c4aa8302bf0c2520bebcd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5321,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=45;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5321,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1478.88,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-03-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a31eefe8567a5bd94f67","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8e2f84e67339ce03369de046","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1459.74","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d905954a49d31c77f02d6fd7ba4de463216275f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5328,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=52;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5328,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1459.74,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-63bcce2c3e45dc2c1a9b","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8b929557929701e15b32aea0","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1516.65","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bdbc45bf378bd37e1da2b6fd853e3e023c0ade94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5309,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=33;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5309,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1516.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-08-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e45206e3d4167f1161a6","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9ce067bc33b726849b4262c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1594.78","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30811510f4c0fd89fed79a1dfc3d83c480663fff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5286,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=10;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5286,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1594.78,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-webdev_arena_external","benchmarkName":null,"benchmarkVersion":"2026-08-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-54609cc4bf6282e2f162","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-494da694f7fb89e7435daf28","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.8 Max.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.8 Max.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Arena Score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_unknown","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id","unit_inferred_from_column"],"rank":null,"rawValue":"1668.94","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d4c93d3bad3d8dcf8c502537712cad5a4ba90d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5280,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"webdev_arena_external.csv:row=4;column=Arena Score","sourceMetricId":"Arena Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5280,"metadata":{"epoch_file":"webdev_arena_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"elo","value":1668.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-60025a15238c1ea628b1","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-70a51f6f6461f3e5d94946d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.41630000000000006","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c36f2d635a7f6e08677ce788f2d520e7868b3116","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5493,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=100;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5493,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-08-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-e26fc78c45af8f103d67","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-273b8cf64b7eb73fcc09af3b","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V3.1","modelRef":"DeepSeek-V3.1","observedAt":"2025-08-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.375","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d760e3ab7092e2e0d23449bc114b8c056cf92102","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5517,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=124;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5517,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-08-21","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-08-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0c6ebc0306a34bf60215","canonicalModelId":"deepseek/v3.1@2025-08-21","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-949888bd402c6dedab3ae8ab","mappingCandidates":["deepseek/v3.1@2025-08-21"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V3.1","modelRef":"DeepSeek-V3.1_thinking","observedAt":"2025-08-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3836","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d760e3ab7092e2e0d23449bc114b8c056cf92102","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5511,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=118;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5511,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-08-21","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.36,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4b327e0b73ba19cb76c6","canonicalModelId":"deepseek/v3.2-exp@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30aeea61146e62cf278af6d1","mappingCandidates":["deepseek/v3.2-exp@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V3.2 Exp","modelRef":"DeepSeek-V3.2-Exp_thinking","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3948","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4273d8415e2553afc4dae6892a9276fdc70c2274","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5503,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=110;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5503,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2025-09-29","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.48,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-18","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f309ac817e7d5e09135b","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ea0616a8cf3375a6e169c9f4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"MiniMax M2.7","modelRef":"MiniMax-M2.7","observedAt":"2026-03-18","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.36950000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0589a55ec9ffef58342599dca79436790f71e7e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5521,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=128;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5521,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"MiniMax","release_date":"2026-03-18","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-01bee4bc7a508a997e10","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-97ba5dd6e240d541b2fd74c7","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Epoch presentation of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4117","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"913f06b049f2c9461de00dab950a736fce96b052","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5494,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=101;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5494,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Alibaba","release_date":"2025-07-31","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1b00caa88997b168a092","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-74961a82b346bea89c0b39e5","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8785","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f724c94a7f7031f08ffd2854044d8e68e678cf27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5400,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=7;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5400,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-06-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fba624f826575d8f30db","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f969159ceef495b3d34ad7b","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9194","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f724c94a7f7031f08ffd2854044d8e68e678cf27","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5395,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=2;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5395,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.94,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c741d0294b314c7bd31f","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b24157f15b2ea85f16c1673e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.454","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ff2c8344d9a71143ad21d3e4c63c1866fe9e54e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5478,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=85;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5478,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-10-15","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0fb380e2ecfd98c76c35","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-26c6d56dc7e55f669ba0bf25","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_16K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4415","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ff2c8344d9a71143ad21d3e4c63c1866fe9e54e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5479,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=86;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5479,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-05-22","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d7a428308c5a228ef756","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-27ed8fd3c63abbb190f4e26d","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4372","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3fc668482f2629c86d27e3732f0f0a750484d4f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5482,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=89;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5482,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.72,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-11-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1cd89970a9fbcec374f0","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cafe8ac867cdaac8e045ded9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6374000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5cc985b29cadc98f210f115fda1cc7c16c5ed46","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5430,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=37;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5430,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":63.74000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-34bf96b5fb76927719e5","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9cc464de77a77276dcd4bb0","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6587000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37de237be1e03faadac914a74b82d20363d28cf3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5429,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=36;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5429,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":65.87,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-50330e349c990d4d6234","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-85f8b7d832cbc59c3dd4f93b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7795000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37de237be1e03faadac914a74b82d20363d28cf3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5409,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=16;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5409,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5721e769f32264ec20f1","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0cc5de6376dbcfd17688bbd3","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_unknown","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.779","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37de237be1e03faadac914a74b82d20363d28cf3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5410,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=17;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5410,"metadata":{"epoch_file":"weirdml_external.csv","notes":"Adaptive thinking","organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-715f5dd212c906c4fe23","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a4dc12ddf962149aa8e22af0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7640000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e2ac81a0f5cea6371336e4fb3077ed4ea73efa0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5414,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=21;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5414,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-3c40eae3e51cc8640ce8","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77431d193ee3b61dcd73ae53","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_high","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7644","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e2ac81a0f5cea6371336e4fb3077ed4ea73efa0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5413,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=20;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5413,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b29cf7ef0b257b5901b0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8043bef1121db11ff2a3b4ac","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7545000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e2ac81a0f5cea6371336e4fb3077ed4ea73efa0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5417,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=24;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5417,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":75.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1b8825a4025859e99028","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-714a7aa316b1adfc8ad0d030","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.764","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e2ac81a0f5cea6371336e4fb3077ed4ea73efa0","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5415,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=22;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5415,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b3a1477bbfbfd44c3e8a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed5a01ca6542e67a88d49c20","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_medium","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7604000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8555156be6579703bb548704b3d58cacf4f315d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5416,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=23;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5416,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.04,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fc3dfe352f55191b8366","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cbb6694a37c21c08de653f90","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7045","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8555156be6579703bb548704b3d58cacf4f315d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5420,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=27;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5420,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":70.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-05-28","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a2c5eef0575f3a0ccf5","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fbdbb17c31e821b4ab1b3621","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_xhigh","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8289","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8555156be6579703bb548704b3d58cacf4f315d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5405,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=12;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5405,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.89,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b6afe3c2ac9c406ad195","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d005807c0b03490f68e4cbf4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8634000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aec74ce5ac9e444e8bd602ca62a1b54ec10b7996","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5402,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=9;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5402,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.34,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-7ff53203d5a8f79a6fc3","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80414d34671bd45745007156","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_high","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9159","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aec74ce5ac9e444e8bd602ca62a1b54ec10b7996","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5397,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=4;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5397,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.59,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f8a5a7fd3ec4a877a053","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-55e653ae540f253d4a4b15c3","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.9178000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aec74ce5ac9e444e8bd602ca62a1b54ec10b7996","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5396,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=3;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5396,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.78,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ffec9cc0e6e950b818f1","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-47db27bb054e48c7e4c8d03e","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.467","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce97e87f3d926263f8edcab0839a582b5373ffd8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5469,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=76;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5469,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-09-29","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b028ce04f9dc3f961588","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0d678187910210d9d8aa3449","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4771","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce97e87f3d926263f8edcab0839a582b5373ffd8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5466,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=73;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5466,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.71,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-b8da1cfb09a2ca7911d0","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-28115bc5890fca3360d2207d","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6607","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"624c98470f43090d6c42f438c1cba4b2787c1a3d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5428,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=35;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5428,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-06-30","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4aa308088166c76473bb","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cda4a2b59b7a1562481db81b","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_high","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6878000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb73bf919463d90ca0facb8e2fc849fe41ecdf56","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5423,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=30;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5423,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.78,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-460f6164c004181fb39a","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-259de5a87e31a39eee0cee24","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_high","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.5705","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"731aed5793dc6040ef4154be7b61a49d34dcbf62","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5445,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=52;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5445,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-31","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-788170370f4306bdb8ac","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-292d44cf56c3a7a1f8a22b1f","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6296","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"731aed5793dc6040ef4154be7b61a49d34dcbf62","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5432,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=39;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5432,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":62.96000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-665f3b779c5e4790d2b3","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8186a8d774b191fbe1d4e7a0","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.4376","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"731aed5793dc6040ef4154be7b61a49d34dcbf62","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5481,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=88;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5481,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.76,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a0c938050ffde5a571ba","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2e06de300e56dca390192e5","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.45630000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"731aed5793dc6040ef4154be7b61a49d34dcbf62","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5476,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=83;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5476,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.63,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-08-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a68e3f24a9ce17ef476c","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e5ad11e3db1c1b66bd869c6b","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6622","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22b9cc1c8e7a15289079a920d0e750f2eeed2ede","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5427,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=34;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5427,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-18659ac27362b94d1e8a","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a3ee02cc2c216d270f498c16","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.46530000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0745f0e157b05f946eac51a966cf0b78168d1b3b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5470,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=77;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5470,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.53,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-baf1ea5beccd91a00701","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-23c98c88f008262185d194bd","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.489","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0745f0e157b05f946eac51a966cf0b78168d1b3b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5462,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=69;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5462,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-0e331c70519936a7377d","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a6eddd4add5a3aaff55dd0a3","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7206999999999999","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba855a12c26d1f23992e784caa4466573d4a31b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5419,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=26;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5419,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":72.07,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2702f53a3b7fefb34bcb","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-603e9e0eafa80fd0497fcbf9","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.39","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4104738bf672a6510136cb234b2446a59ab8493c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5508,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=115;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5508,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-05-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-186f32ee066fd35e694c","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-21b0bf39edec2b397e38163a","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6264000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a6824511a7174e1693d2daad79ae3ea00024cdc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5433,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=40;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5433,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":62.64000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-21","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-06700b0983eb366f2cef","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2c998902ee64a771aa36c94f","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.561","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4edd2a6e1711b46b1542f850bb2e7e4600b56bfa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5446,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=53;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5446,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.10000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ff18e74943de1db103af","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-795d056d0612c985da07188a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4817","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5a678758e5a861c7524faebedd138ed85ee351f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5465,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=72;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5465,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fcaf62f6cad9f4deee58","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3680011d4687b16ad4d4bec7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5710000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13d535499c350d1714d23fac6014ed99d8c5dd9f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5444,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=51;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5444,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.10000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8ac52a43361869614c9f","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3c49008818d7c650e2d88f69","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.2","modelRef":"glm-5.2_high","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6731","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a835e18da0b86d073ef880812fd4a1bf8344f2d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5424,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=31;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5424,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":67.31,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-06-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d09da0eb7c80fa17ae34","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4e774ba1ed1884cfc1bc5089","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7012","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a835e18da0b86d073ef880812fd4a1bf8344f2d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5421,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=28;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5421,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":70.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-79be2a9df6db0bc2b3d9","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f08ecad1a9135a58141123e5","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6070000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44cf260e2654b7f5e64115459f54f20d34199e5e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5437,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=44;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5437,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.70000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ebd76135080039ad81f2","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ac3c3200b4989717b6b0918","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5267000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f54a1227fe9648fe02a76fc92b079478d0172fb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5452,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=59;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5452,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":52.67000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-d003f6ab2b199397f37c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1dfd18e66bbec76879f41c66","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.38060000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b44ff24343e5f7a4340f305a431e9b50ab72bf32","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5512,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=119;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5512,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-08-07","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-c73bce496b003f6a1fb9","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2e7a30afbdc6fed07d0b1c90","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.2589","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b44ff24343e5f7a4340f305a431e9b50ab72bf32","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5538,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=145;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5538,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.89,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-11-13","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-948a266e35edb025d087","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-46c95a1cb2b5044696514e74","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6077","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8608389901c7d1d98e8b5b04f100e210bce20a70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5436,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=43;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5436,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.77,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fca5e810b6442e4e0422","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7d107d2df1f1a7b2074e11dc","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.496","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d07881215fbf18f6aa4c1f0f1ef0a8f54011c07","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5460,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=67;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5460,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-16e768f31b037c284811","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77a5a2cda57aa2c67af5e72f","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6344","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d07881215fbf18f6aa4c1f0f1ef0a8f54011c07","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5431,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=38;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5431,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":63.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-684b4c29c00c86805dca","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4a7df37769a5ffeb7a87a81c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.49600000000000005","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d07881215fbf18f6aa4c1f0f1ef0a8f54011c07","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5459,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=66;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5459,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.60000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-12-11","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-00fe9c469d5b277ee0c0","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8339942025fa093de98db7e0","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.7219","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d07881215fbf18f6aa4c1f0f1ef0a8f54011c07","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5418,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=25;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5418,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":72.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-1d026a2c5c66ecd246e5","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33355cc4465735c14669be60","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.793","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d872a32641531e6593f511a3bc39aa726b520683","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5407,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=14;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5407,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":79.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2ef84f382dab1ec1cc63","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8be5de2765d1a0ee41911cb4","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex_xhigh","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.779","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d872a32641531e6593f511a3bc39aa726b520683","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5411,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=18;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5411,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-fa7d30e9d3bddf46fbca","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-888fd7e64d89b4d91f17a03c","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5744","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"702b7082f5727dc39d94e27b1deac56bba85bb5d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5442,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=49;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5442,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.440000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-96f2050719c101ac3c75","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a48d641af6303c46c86c6563","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.777","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"702b7082f5727dc39d94e27b1deac56bba85bb5d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5412,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=19;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5412,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-199168bb7d93905328aa","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1516370a76499e4fc263ff8a","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.603","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6be839fc4954a4661d3f1cc1d292ff62a9e1f771","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5439,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=46;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5439,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6ce5591b46a672c86ca1","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f7cc55c6ebb2c5ccf67374c7","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.379","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6be839fc4954a4661d3f1cc1d292ff62a9e1f771","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5515,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=122;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5515,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5a68d66a68b73c9c3e30","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f890ffd1cc6393f9f11129d","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.49229999999999996","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0934bbde1993dd9841f8024f37763ea4bc7bfba1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5461,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=68;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5461,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.23,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-17","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4470fef6e48707165028","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0a83dca0a66c1875a63217f1","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.3796","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0934bbde1993dd9841f8024f37763ea4bc7bfba1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5514,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=121;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5514,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.96,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-03-05","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6b57f88ff29a7339e5a6","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8fee6d78fdbf8b55b06c2249","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5744","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"246c340f29708f931346f1e70e7edf59b8a244d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5443,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=50;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5443,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.440000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-f6b83cf5f499ada80f61","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-83fb5797e29ed6eebbcc711c","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_high","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8390000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70da2c567f7fa22622061aac395ca96a7e7ef0c1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5404,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=11;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5404,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-ce6b0463e340258a0653","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b54b7b5f9d4a33740b1d4c05","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6715000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70da2c567f7fa22622061aac395ca96a7e7ef0c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5426,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=33;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5426,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":67.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-23","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-25ee75ab9da1938c788b","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9c988f47aac758fafd45569e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.5","modelRef":"gpt-5.5_xhigh","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8491","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70da2c567f7fa22622061aac395ca96a7e7ef0c1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5403,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=10;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5403,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.91,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-a8152960eba588e745d7","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d169fe8b2b60050fc5689be9","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.6086","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04e146e9ca2c3251ae371318b487da6b33826769","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5435,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=42;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5435,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-089f8448bda2b5844612","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-96741dc153c02e43f48d03ae","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8876000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e68a24de79e121be3ee44208b1db3aec111b3cb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5399,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=6;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5399,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.76,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8e621cb54fdbf780c49c","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-65a29c0e79942adc09cbda5b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8697","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e68a24de79e121be3ee44208b1db3aec111b3cb","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5401,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=8;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5401,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.97,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-62f228fb90a89d1a3e9d","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b912ba52134065df48994e7c","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch Pro-compute presentations of GPT-5.6 Sol; the compute setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_promax","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.8943000000000001","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e68a24de79e121be3ee44208b1db3aec111b3cb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5398,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=5;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5398,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":89.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-09","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5f5d86bdbdec28783205","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-858cdf9a82f6b5f1a6eb9ddf","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_high","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.7827","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3183873d85130fe3fc4a8437ab46be822115468e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5408,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=15;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5408,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.27,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2025-09-19","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-2cfef1ddbd7925f583a6","canonicalModelId":"xai/grok-4-fast@2025-09-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2bcdec59010ff9468cb634c9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Grok 4 Fast","modelRef":"grok-4-fast","observedAt":"2025-09-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fe549410628c112f83767474afea46eecd5f7b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5487,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=94;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5487,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"xAI","release_date":"2025-09-19","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-08","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-6652a159c8d045ae7a27","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-341ed094bfdff0baa231e202","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Grok 4.5","modelRef":"grok-4.5_unknown","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.4643","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6d91c8cd27fe445c5b2370fec6e83698a60f3cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5471,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=78;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5471,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.43,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-08-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-590ceecbaeede2ad000d","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f347c0baceb90614d1387b80","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.6729","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d32ee225f22ad9a648e19894183d3b65d6959b90","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5425,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=32;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5425,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":67.29,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-01-27","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-88098697203fcfcc9f43","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://htihle.github.io/weirdml.html","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f473c1a905dcf870884661af","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.456","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"750e291ff8cf1b56568cfe07e185511a41f0f2bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5477,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=84;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5477,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":"https://htihle.github.io/weirdml.html"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":45.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-20","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-bf45e50c7bb726ba6d71","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5bfa42cd5b40bfa478f29f04","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63f255f02cbcfa44914063b2797c723e7ed614ee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5447,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=54;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5447,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-06-12","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-54e123589fe1b6b9f917","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-756a44a8c91a509d54363847","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.5412","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85160d8d46eac760b94c55535516afd3799d804d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5450,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=57;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5450,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.120000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-07-16","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-5793df6abb736edbc41b","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-020f124c2c59b561943f7f81","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"0.8257","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fe16409b2fe8c1da14533b4ddf234f7528858dbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5406,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=13;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5406,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.57,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-02","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-9f596582554d7a9c7b64","canonicalModelId":"qwen/qwen3-coder-next@2026-02-03","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a9cc690a7dc306eb133c5521","mappingCandidates":["qwen/qwen3-coder-next@2026-02-03"],"mappingEvidence":{"confidence":"high","note":"Epoch exact Qwen3-Coder-Next spelling.","registry":"model_aliases.json"},"mappingNote":"Epoch exact Qwen3-Coder-Next spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3-Coder-Next","modelRef":"qwen3-coder-next","observedAt":"2026-02-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.344","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69877805ac133f171a811b2a9cb618ff33c621d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5532,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=139;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5532,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"","release_date":"2026-02-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-02-24","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-4b19643c6f48b71fac38","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-be0a409296418777aa0b278f","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.5 27B","modelRef":"qwen3.5-27B","observedAt":"2026-02-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.39530000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"575937384ed137d817321f88e312b0c8afc6a7c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5502,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=109;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5502,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.53,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"epoch-weirdml_external","benchmarkName":null,"benchmarkVersion":"2026-04-14","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-src-epoch-benchmark-hub-8ee99487ab649c94d0cc","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1077e16007cef6d06f4613e8","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","source_version_label_without_id"],"rank":null,"rawValue":"0.34490000000000004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"337d4bccfe8e9d8a69cb7f530950440157b50d0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5531,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"weirdml_external.csv:row=138;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":5531,"metadata":{"epoch_file":"weirdml_external.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.49,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3434e7bac5483b60d9fe","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b321d60b78f3ea62eded4c12","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041379310344827586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f69ea38bf7feb879ed9fc7937c3d29974ce2fa3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2457,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=65;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2457,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.137931034482759,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2ce7801871679556cbce","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-16278bc1a4f0103af49e5c8e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.059027777777777776","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f69ea38bf7feb879ed9fc7937c3d29974ce2fa3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2456,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=64;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2456,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.902777777777778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-60cf38713ef657d080f0","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f1f3e6d382f1e8987775d314","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.04482758620689655","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1868c8159bdec36b207646b5b283b5a5a88eac84","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2462,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=70;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2462,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.482758620689655,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5f068783a3bbda54b454","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f9800e442d92c647a386184a","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_27K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.041379310344827586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1868c8159bdec36b207646b5b283b5a5a88eac84","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2459,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=67;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2459,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.137931034482759,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-108a2a69a8cb3ffe8794","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-244fab2c69780ad856124bbb","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.20689655172413793","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb8c1be6ecb94e8453203bd952848d16c4af5cf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2433,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2433,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.689655172413794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f18c5e443a690ad95bd8","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c009d95d5bd9f4e460a8abf2","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.20344827586206896","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb8c1be6ecb94e8453203bd952848d16c4af5cf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2435,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=43;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2435,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.344827586206897,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3694586aded7b91c95b7","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fd0f5b1796243a99e2c6913a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.20689655172413793","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb8c1be6ecb94e8453203bd952848d16c4af5cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2434,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2434,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.689655172413794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2be8acf006d07c8018e9","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3f85d87c2fb453b4ccb73b20","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.38275862068965516","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"781251315f12c51e6b24ff2485965c3f55b28519","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2418,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=26;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2418,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.275862068965516,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3808c86227685ae0108b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-179e287821fc3b106aa1f568","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"781251315f12c51e6b24ff2485965c3f55b28519","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2417,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2417,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4508fbeca11f4dbf469f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3218a0b0944ad986a30e2c1","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_64K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.39655172413793105","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"781251315f12c51e6b24ff2485965c3f55b28519","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2416,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=24;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2416,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.6551724137931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-921c5508106840f441e6","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b76a51b69fb2f42d973d56cc","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"781251315f12c51e6b24ff2485965c3f55b28519","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2415,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=23;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2415,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b5b2d4a388b95c2f3136","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3d26ba3120baa77358baf6b9","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4379310344827586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"304a25b8cf1b0cbc31f270487eb52b8ad831028a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2406,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=14;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2406,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.79310344827586,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ea32fdf86850c7ef7de7","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8de509232aee018c87ee66cb","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4724137931034483","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7719495619269849890d07a1436ebc26c5cd8e4e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2394,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2394,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.241379310344826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d639fcfbdf7ff1f49921","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c5e057d430672fc0d39ca03f","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.09310344827586207","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76dc0438aaec64d5ee8e663ed69d3d083c6e8f53","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2442,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=50;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2442,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":9.310344827586208,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-151e239a049a00e57339","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aa1e8961b2db1d9450869636","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.1522491349480969","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76dc0438aaec64d5ee8e663ed69d3d083c6e8f53","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2443,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2443,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":15.22491349480969,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-70bd4351c734d9058d32","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3651a98edba24e0434e649e3","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_59K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.13494809688581316","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76dc0438aaec64d5ee8e663ed69d3d083c6e8f53","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2448,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=56;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2448,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.494809688581316,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8eeca0f54c38073535cc","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-71b34ec3865f180c97542368","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingEvidence":{"confidence":"high","note":"Same Sonnet 4.6 release with effort/context/search presentation.","registry":"model_aliases.json"},"mappingNote":"Same Sonnet 4.6 release with effort/context/search presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_16K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.324","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ff30dc8cd0e4b790c1fa62046fa79cccf5b0f51c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2412,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2412,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-12-01","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-14aa9ac87019dffa4e73","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bff9ae4f39c576cf377ba322","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of DeepSeek V3.2.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of DeepSeek V3.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V3.2","modelRef":"fireworks/deepseek-v3p2","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.221","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a475bb0378cb6ff7042ace21853e53c61af55674","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2421,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=29;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2421,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-859a1064dfba999fdb5f","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-19d061b04065c0aa73992268","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.279","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fe27fc59eba77e73f5eb47e2d1d139e3ae6b1ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2419,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=27;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2419,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.900000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0d5ec329baea26840149","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-50ef7ef9e24438145e5d0824","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Flash","modelRef":"gemini-2.5-flash","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.04844290657439446","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6646ffcdeb2db37a71cff756117498db15ab728d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2422,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2422,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":4.844290657439446,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dce57df924f38b1bd60a","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b009fdc29c1e261afbc4bd51","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.1413793103448276","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7567207db96c50567dcf6b2613eff2bcfbe3eae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2436,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=44;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2436,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.13793103448276,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3f8df35d228f0ddb2f72","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2a2a927857bef3d0f29bc87e","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.369","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b230fd128e62d4ffed43475ba7d1caf2dc459d98","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2414,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=22;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2414,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7ebb4b2162a7f53f2e80","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f76ab8b45ad0fcbd4bfa9b3a","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3896551724137931","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"671040bc54d1587c079316913625487033199606","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2396,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2396,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.96551724137931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-11","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-673849b7dae070569643","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6972409a2ce270b48e96b0c1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.16433566433566432","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f1f81c72e9b903b15e143ef19b4d06875c9b2ec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2413,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=21;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2413,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":16.433566433566433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-04-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e39ed23304f7ca5ebb4f","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-172dd16f4db4a31bee79a406","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.33448275862068966","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f151db736337db4a74da5fa85d7ee174049ea922","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2401,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2401,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.44827586206897,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a0d825a32b46e3f54920","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-691c64ddb49bef3dc9b49fc8","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.32413793103448274","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c6306a64cbae9a14435a3c881fdd5870f537baf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2447,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=55;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2447,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":32.41379310344827,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-10bcf7ed9cdce990d8ad","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b50c5abbcbd2c7de889c58eb","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.27241379310344827","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c6306a64cbae9a14435a3c881fdd5870f537baf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2454,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=62;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2454,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.241379310344826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-95721566d60e9bb05b18","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1c2caf9cad755c097950f98c","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.27241379310344827","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3a0a48f9d613071a8f8447499a4a63a649b707f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2449,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=57;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2449,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.241379310344826,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-01cfef6528083b8c5843","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c66063457db5bab8b413d306","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.20344827586206896","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3a0a48f9d613071a8f8447499a4a63a649b707f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2452,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=60;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2452,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":20.344827586206897,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0edb9797156d24276815","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ba7029ce10b2ab5338bb1bd4","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.08275862068965517","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f440f303a373851737389acb610e19689ba7340e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2455,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=63;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2455,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":8.275862068965518,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6043a3ac3f6af07fff6d","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6ec7c2838844c157df7c2d5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.07241379310344828","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f440f303a373851737389acb610e19689ba7340e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2458,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=66;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2458,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.241379310344828,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4de4ae8bf3486601fe1f","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9c536a67d10b060556a6c65f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3103448275862069","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6fc7c37da8fe75d0fa94b961865cf2318ef5061","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2446,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=54;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2446,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.03448275862069,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-972ed165388576b73191","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b85ec19119b0ce9fa34de18a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_low","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.17301038062283736","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6fc7c37da8fe75d0fa94b961865cf2318ef5061","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2432,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=40;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2432,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":17.301038062283737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e175cfb98de028b73c24","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-790020830060d671f0a82857","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2689655172413793","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6fc7c37da8fe75d0fa94b961865cf2318ef5061","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2438,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=46;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2438,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.89655172413793,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bfab50f54775514e77a6","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0b0bf2860cad8a0256fdd69e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.02097902097902098","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6fc7c37da8fe75d0fa94b961865cf2318ef5061","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2431,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=39;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2431,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.097902097902098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7a672c28d1411cffb0ee","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-071819aeeed98d8c7d2821bd","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.40344827586206894","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fba1122bd3459b63e40c373ab9bd03c8b010112e","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2425,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2425,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.3448275862069,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b369668950a765856428","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a8c9cc84832d82599340a295","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2655172413793103","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fba1122bd3459b63e40c373ab9bd03c8b010112e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2427,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2427,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.551724137931032,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-29d4397f465bd749dc9a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d1152a30a69a82f52ad85a68","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3689655172413793","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fba1122bd3459b63e40c373ab9bd03c8b010112e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2426,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2426,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.896551724137936,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-27f33eb2f1f2d7147151","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bda1f9093a27d85ed88f4fce","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.407","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fba1122bd3459b63e40c373ab9bd03c8b010112e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2424,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2424,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bab6d7a70dcc6643b09b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5f9112587dbb0c00315cad64","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.476","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de087ff82e83c2ee80a53b39bafa64149a6eb0c0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2410,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2410,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.599999999999994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8f5978fb415be38a1df2","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-18d300317eb4909b9b3f6593","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2828","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b0ed81f75d61c53a59d535f4d6435177be24d32f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2407,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2407,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.28,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4cdd44789ed684f82d73","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c0297a540201a8f071c6488f","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"592e04f4c9e2690197c22bd572fa2067d245821e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2408,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=16;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2408,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.86,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b60c03a6463d6d976d3a","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-14612a3c3f64ab64aa677261","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50abe0065724c33cfa646ae089f2fe0bacf72640","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2411,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2411,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-db9a95d0952788654eb9","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33827105735abbf721133218","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3896551724137931","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fb4575b39cfaa20df9f84acebc8604dae79fc70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2402,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2402,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.96551724137931,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-25","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-31f95f319261a9fbde32","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cae9720fdc1e866d14a0165a","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.06206896551724138","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4ecba29ef662bfae52848f2bdb027a469bc6a9c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2400,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2400,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.206896551724138,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-02-16","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-17abb51de0d95df21dee","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2c5096d3e72a21c1a3349a16","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2103448275862069","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e261d1410608612f483efb17f7e745bc3f9a7ffa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2397,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=5;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2397,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.03448275862069,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-04-27","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c68a4ebbd52dbf047a03","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b46a82075d9ce4c0b83c2613","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.10344827586206896","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63134279029f7f75756f97bf2b30628d2996d665","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2399,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2399,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.344827586206897,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-98cc53b9f7353a8704b6","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-328c33c5e10bcbbbb29a46d7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.23103448275862068","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60a36465868bf70c9a8d9bf072a8f3c5a1733e43","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2395,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2395,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.103448275862068,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2026-03-31","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6c4862c56ecf464769ce","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98d867e21180d8bc37fd3248","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2620689655172414","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8964fadcacfad234a3fa0ed719e240ceac682c2d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2398,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2398,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.20689655172414,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"frontiermath","benchmarkName":"FrontierMath T1–T3","benchmarkVersion":"2025-12-22","benchmarkVersionHint":"frontiermath@v2-t1-t3","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4138140a651792f2d39f","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-deb3761ae298f4785b0bbd76","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.024390243902439025","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d08aee11b2b80dcd398bfd1dab7f8c3d610c4c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2420,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"frontiermath.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2420,"metadata":{"epoch_file":"frontiermath.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":2.4390243902439024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-05-28","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c592f8e8f54939299bdb","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4fe3042119355255cc9c11e5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7632575757575758","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5748c32fb712b343c7546352a88b6de93746f2e7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2834,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=160;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2834,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.32575757575758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-01","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-89fb84bee545f13b1ec0","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-921524d4f34b011ec22e943e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"MiniMax M3","modelRef":"MiniMax-M3","observedAt":"2026-06-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9090909090909091","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"727c1ba1df3344b9d41fa3afa33da3e562cb8336","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2682,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2682,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"MiniMax","release_date":"2026-06-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.9090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-8cf21e60780841503421","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/08/7b/087b0dc6-3c87-4f40-8458-16f19de92200.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-8b6b9e7f732e37963801eb54","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.1","modelRef":"MiniMaxAI/MiniMax-M2.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":63,"rawValue":80.8080808081,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bf0df3fecc28d678295d467b4a38c7580a946e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":63,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=63;row=62","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":63,"metadata":{"author":"MiniMaxAI","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":37,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/08/7b/087b0dc6-3c87-4f40-8458-16f19de92200.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.8080808081,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-ebe56f1f1ea84f5cbf00","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a7e89b1f4e9a5fbe4fa482a2","mappingCandidates":["minimax/m2.5@2026-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.5","modelRef":"MiniMaxAI/MiniMax-M2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":41,"rawValue":85.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2698e516df5393ab31d552cf21b5b514c966c6b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":41,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=41;row=40","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":41,"metadata":{"author":"MiniMaxAI","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":228703644928,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-f599eca64a499a94f90e","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-549c41eb2cf8950f7e9eb3eb","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa_1.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":29,"rawValue":86.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3512869a68494efb377147f570c494647cb0adb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":29,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=29;row=28","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":29,"metadata":{"author":"Qwen","filename":".eval_results/gpqa_1.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":8,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-edeee05f5492a2006bbc","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/58/60/5860f5bd-926e-4ece-bc87-a1af551a6280.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-01452c192b3b2c2863a44824","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":43,"rawValue":84.8484848485,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3512869a68494efb377147f570c494647cb0adb3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":43,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=43;row=42","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":43,"metadata":{"author":"Qwen","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":24,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/58/60/5860f5bd-926e-4ece-bc87-a1af551a6280.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.8484848485,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-afa775eeca3b6dd248d3","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-ebf26d21d1bac77bf23ba7e7","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa_1.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":39,"rawValue":85.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"31bd84db1e8b785f862f9044d4d3f2b0ba7a2eda","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":39,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=39;row=38","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":39,"metadata":{"author":"Qwen","filename":".eval_results/gpqa_1.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-cf44668f728225da6ca8","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/5b/a4/5ba4cf19-18b3-4d08-bad0-ef29422cb18a.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-1787df51008565c9f6882080","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":59,"rawValue":81.8181818182,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"31bd84db1e8b785f862f9044d4d3f2b0ba7a2eda","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":59,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=59;row=58","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":59,"metadata":{"author":"Qwen","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":47,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/5b/a4/5ba4cf19-18b3-4d08-bad0-ef29422cb18a.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.8181818182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-d14d2075ca5ef12f785a","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-9b5b1e46f0a27386bda9717a","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa_diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":15,"rawValue":88.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c72922fa20a4f22b394595fa144eab0bbc97d1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":15,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=15;row=14","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":15,"metadata":{"author":"Qwen","filename":".eval_results/gpqa_diamond.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":9,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-43c6d0cac418445c58f1","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-ef8f15427e7456b63f39ff6b","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":22,"rawValue":87.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85c09cfd886acfe895f7c7c4f41617f43c726e11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":22,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=22;row=21","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":22,"metadata":{"author":"Qwen","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-dd4cbbc9a83331c1976a","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4e3425007ca54f34dddb8ac3","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":35,"rawValue":86,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"014e01a58cc243885e7ed630b3c5da1cc64aa530","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":35,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=35;row=34","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":35,"metadata":{"author":"Qwen","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-b5e785874624593e4f45","canonicalModelId":"qwen/qwen3.8-2.4t-a95b@2026-08-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2cb08742678a21b103739b81","mappingCandidates":["qwen/qwen3.8-2.4t-a95b@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 2.4T A95B","modelRef":"Qwen/Qwen3.8-2.4T-A95B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":3,"rawValue":92.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2989a7d1b6533da666e4e5b4508636ce0e0a5641","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":3,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=3;row=2","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":3,"metadata":{"author":"Qwen","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":2446182725504,"pull_request":6,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-e1ad0278a51c3958e6f4","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-6919e4db816f8d64a713bc5a","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"Qwen/Qwen3.8-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/Qwen3.8-27B.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":89.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae9639466975ab97c9f82d3b42c150e7b9a7b011","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":13,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":13,"metadata":{"author":"Qwen","filename":".eval_results/Qwen3.8-27B.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":22,"source_name":"Qwen3.8-27B model card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-c5130237440ac554d7ab","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-5933de1179e8edb153efdfc6","mappingCandidates":["qwen/qwen3.8-flash-next@2026-08-26"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"Qwen/Qwen3.8-Flash-Next","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/Qwen3.8-Flash-Next.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":4,"rawValue":91.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c8da4b3b7a3205b405d89c63bf04c3443df6db9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":4,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=4;row=3","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":4,"metadata":{"author":"Qwen","filename":".eval_results/Qwen3.8-Flash-Next.yaml","lower_is_better":false,"num_parameters":179999981459,"pull_request":12,"source_name":"Qwen3.8-Flash-Next model card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-282f9c65b6385fe9c6c2","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/36/22/3622b785-41c0-4a96-b867-5bde46851587.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-872a1f364f5dece8e63f2e63","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo V2 Flash","modelRef":"XiaomiMiMo/MiMo-V2-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":52,"rawValue":83.3333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcfd8dc9b5ca8a8ae3a84a6b933057da93eed639","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":52,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=52;row=51","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":52,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":309785318400,"pull_request":42,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/36/22/3622b785-41c0-4a96-b867-5bde46851587.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":83.3333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-acd733784e53f80ae983","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-669d356be453b229ccde054d","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5-Pro","modelRef":"XiaomiMiMo/MiMo-V2.5-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":93,"rawValue":66.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cebbce23277718ead2e8b242fdc32c9878087236","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":93,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=93;row=92","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":93,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":1023244718976,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c336ee5aa3b1dc9ccdd4","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ef5d1f3a7ccef5947cf20e89","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_high","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3286be7e8bb21745748a1a349e2f3bb6b98617a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2744,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=70;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2744,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1327138a40d7a86da7fc","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eeeb81aa467b0f4ca561cd52","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_low","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7878787878787878","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3286be7e8bb21745748a1a349e2f3bb6b98617a9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2743,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=69;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2743,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.78787878787878,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4bc46055e66839fca014","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e8c100081b9dc12e6eacc395","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_max","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8585858585858586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3286be7e8bb21745748a1a349e2f3bb6b98617a9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2742,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=68;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2742,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.85858585858585,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2418296368f55245de2c","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ce414f86948ba83d3b9803d8","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6047979797979798","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47e5140df3cb15889fc16c9af11f35cbdb20ccaf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2822,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=148;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2822,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":60.47979797979798,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c89f133e2cb6b2840538","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64b456bd45990904d2c68569","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7121212121212122","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"47e5140df3cb15889fc16c9af11f35cbdb20ccaf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2820,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=146;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2820,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.21212121212122,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-50a759c40a31cc9715fa","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d3815943035ba8563ee37a12","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6919191919191919","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f96ab76db602fd839a46dccb0bcfc526c53a204","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2844,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=170;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2844,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":69.1919191919192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fc1068769baebd5ee44a","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4ff5876157472d164455c430","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_16K","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7626262626262627","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f96ab76db602fd839a46dccb0bcfc526c53a204","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2841,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=167;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2841,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.26262626262627,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c7c81c440cdfdc55e2f4","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a0b516b6a34008903657994","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8068181818181818","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ddfb3969f6d273d151ba7eee31c2acb662604a32","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2808,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=134;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2808,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.68181818181817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dbbe20c2428c37bae7ba","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c71134eb80c7ee0f3bd513a7","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_16K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8547979797979798","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ddfb3969f6d273d151ba7eee31c2acb662604a32","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2806,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=132;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2806,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.47979797979798,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1a7f4ee36c065f8b6b42","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d30f283391a157755bf30ae1","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.860479797979798","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ddfb3969f6d273d151ba7eee31c2acb662604a32","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2807,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=133;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2807,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.0479797979798,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5c294baf29f3c19114f6","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5f08c08d27f0b5b80e3252fc","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9053030303030303","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86b66dce70a4b1a69b1f6be2ce2afa31081de5d5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2794,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=120;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2794,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.53030303030303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e900696b54601487af21","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-30a8c1a35c9d3e5129cc91f7","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_64K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8876262626262627","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86b66dce70a4b1a69b1f6be2ce2afa31081de5d5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2795,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=121;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2795,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.76262626262627,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-db75c60eae9e377c130e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c6ba0cf59dd68267c2e66dc6","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8838383838383839","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86b66dce70a4b1a69b1f6be2ce2afa31081de5d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2723,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=49;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2723,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.38383838383838,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3112257bc0f92c0d3c38","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f762b8966beca6fd93e61e1c","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8636363636363636","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e38b3113b6b3fab3a5ba3de42a0e0ef285b2381","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2741,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=67;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2741,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.36363636363636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f664b2ce7da964143413","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b444cb3f72cdcc8594d0c6b","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9015151515151515","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e38b3113b6b3fab3a5ba3de42a0e0ef285b2381","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2785,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=111;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2785,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.15151515151516,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-96ac33a4dad4653a8a0b","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dad525f30663a32c8629db86","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_low","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8838383838383839","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25056fb59496e2f31fdcd67343d59df66aeff933","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2740,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=66;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2740,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.38383838383838,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6a558b85ae3345c295f8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e569922d07f2891afa460452","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9103535353535354","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25056fb59496e2f31fdcd67343d59df66aeff933","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2779,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=105;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2779,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.03535353535354,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c91451e8898ea8e43b96","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-010d54fff00ef7226d59b01e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingEvidence":{"confidence":"high","note":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","registry":"model_aliases.json"},"mappingNote":"Epoch capacity/setting suffixes on the Claude Opus 4.8 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_none","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8535353535353535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25056fb59496e2f31fdcd67343d59df66aeff933","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2739,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=65;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2739,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.35353535353535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6907f6ac3a0c29fbbf67","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-de383f1bd8b384fff42a5131","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9292929292929293","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4474e9d591ca3cbd46f5491b1d595a3c0a918981","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2737,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=63;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2737,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":92.92929292929293,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-74046e5923044d0bf5b2","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-42fa8ec29f4bb7e2eef887fe","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_low","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8787878787878788","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4474e9d591ca3cbd46f5491b1d595a3c0a918981","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2738,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=64;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2738,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.87878787878788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d48cb6b91caee0de831b","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-293dc4c10503df5943b1fb9c","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9387626262626263","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4474e9d591ca3cbd46f5491b1d595a3c0a918981","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2752,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=78;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2752,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.87626262626263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f6923cdc1a5ce85235c9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-504da72f32cb3f60cc39bbae","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7373737373737373","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef81c033fc5f54639493df403cbe686b99f6702b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2824,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=150;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2824,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":73.73737373737373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7bcab888f0654eb2b6a9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a2a7ac9b402de73084096006","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_16K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7878787878787878","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef81c033fc5f54639493df403cbe686b99f6702b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2817,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=143;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2817,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.78787878787878,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-99a82f4fbc1b67127a76","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bc1c455b80d920790e7b106c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_32K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.817258883248731","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef81c033fc5f54639493df403cbe686b99f6702b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2821,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=147;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2821,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":81.7258883248731,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b6083bf2a6c0efc21223","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-418291204b36d7c753c7f083","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_59K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8232323232323232","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef81c033fc5f54639493df403cbe686b99f6702b","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2818,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=144;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2818,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.32323232323232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f0913e96c41f6d8b77f5","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-71211b0b70d5a8fbcc93de61","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_32K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8737373737373737","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1db95f8b16e01d9b290147200cc340731dfa0faf","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2791,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=117;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2791,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.37373737373737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7052e8d6b636bb583f1f","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-53f2ce9a365f458875a0c5ad","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_high","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1db95f8b16e01d9b290147200cc340731dfa0faf","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2765,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=91;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2765,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-11e45ec78117bca56778","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-85dd4ef43c66de58a71a5339","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_max","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7878787878787878","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1db95f8b16e01d9b290147200cc340731dfa0faf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2736,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=62;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2736,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.78787878787878,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e78be07775c14aafeb28","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-80ac6b5c90318ee875adfb3d","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_medium","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1db95f8b16e01d9b290147200cc340731dfa0faf","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2768,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=94;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2768,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-30","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-71ccfc0919f4a99bfea6","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-48ff75929bdc171e3f97eae5","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.803030303030303","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e1cbe5291952137ef6074a998ee5a9f8c4840647","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2735,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=61;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2735,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.3030303030303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-30","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9a10f895c6c94f81fec4","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-43e0ac11a075aad69dff57f8","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9053030303030303","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e1cbe5291952137ef6074a998ee5a9f8c4840647","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2775,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=101;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2775,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.53030303030303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-6f96276c42cfea9ff879","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/a1/dd/a1dd2ea2-f092-423e-bc53-04c2fba5948d.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f4b5ce6a8ab5d45cd89a9c64","mappingCandidates":["deepseek/r1@0528"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek-R1-0528","modelRef":"deepseek-ai/DeepSeek-R1-0528","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa_diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":74,"rawValue":78.7878787879,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b680e20b48cc191878ab33c926aade42543db4e3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":74,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=74;row=73","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":74,"metadata":{"author":"deepseek-ai","filename":".eval_results/gpqa_diamond.yaml","lower_is_better":false,"num_parameters":684531386000,"pull_request":117,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/a1/dd/a1dd2ea2-f092-423e-bc53-04c2fba5948d.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.7878787879,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-0933d531c21a0ece008f","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-7f770c6a659f6e05ac3f7519","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-ai/DeepSeek-V4-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":19,"rawValue":88.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c6e105ea8c5e0efb228316fada471b0113643d2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":19,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=19;row=18","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":19,"metadata":{"author":"deepseek-ai","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":290944616402,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-ae94a2e43d0b25c75f92","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-213e986a2f07ced2aef02423","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-ai/DeepSeek-V4-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":10,"rawValue":90.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6c20badbe6ddaa9064b2d62f83cbfda37e0bf12f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":10,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=10;row=9","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":10,"metadata":{"author":"deepseek-ai","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":1598839674782,"pull_request":110,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-01","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cad69d56d11ca71f9d77","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2fb198af14e00fd4f3cab7a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V3.2","modelRef":"deepseek-chat","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7121212121212122","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a7c87a07907d35e0d5fa6d59be40471c2a851374","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2757,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=83;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2757,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2025-12-01","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.21212121212122,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-31","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7c4ef51159bf794fe054","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0ebdfb9a0bed05d3f3e1a8ea","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9103535353535354","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57b9ce96d55ca7cce9e17422f33a4e0d05c35859","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2750,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=76;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2750,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.03535353535354,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7b3633c8798f4826ecd1","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c41eebd48ab17f806c3d3261","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9166666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec7dd6dfa82604c2ccb3df1fc71984da7052e75f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2677,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2677,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.66666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-daf92c819b10d16e59cf","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-93832bafdc60ffb6d2648eef","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_high","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9090909090909091","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b75c1173b6749bba3fc3476aba50aa0f0a783a6b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2722,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=48;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2722,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.9090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3d737aa5e3b15c9a0f2d","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7fda6b6774934380a3289bba","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8964646464646465","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b75c1173b6749bba3fc3476aba50aa0f0a783a6b","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2778,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=104;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2778,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":89.64646464646465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-24","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-76b62cdccbb5ba2e5c28","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a3cc6e82dde967056ac7c83d","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on DeepSeek V4 Pro. Date-bearing 0813 aliases are intentionally not collapsed.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_none","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7323232323232324","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b75c1173b6749bba3fc3476aba50aa0f0a783a6b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2721,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=47;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2721,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":73.23232323232324,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5b4b2d378d952ec25c4c","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0cbac08478a0835c4a67d714","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.876","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8c20f53bcf3db0c5accce7e8405f618e9e9bc48","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2796,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=122;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2796,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c5094ebeb100354f5022","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c3971f1e4fdaf1c9aaee4b9f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8529040404040404","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d6ec12087ad63318adab34d98075c0a5c5960bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2811,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=137;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2811,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.29040404040404,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bef446940b20ea14a190","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aebbfcb47ec31709acb564a3","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.941","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2236bb48654468618ea4578463314e408191b0d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2792,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=118;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2792,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-486689be95429a9ddd55","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-50fac8d3b1963045dc8a786a","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview_high","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9444444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2236bb48654468618ea4578463314e408191b0d7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2730,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=56;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2730,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b282ab430de2c3a58db7","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec5a110adae2297425fa20b2","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fcaf3cb464ac261f79da3caf1d671f1dc95f3050","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2728,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=54;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2728,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e9404e6f1764c0dd47b2","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d72cb63d93dcca012e888fbd","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7575757575757576","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fcaf3cb464ac261f79da3caf1d671f1dc95f3050","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2726,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=52;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2726,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":75.75757575757575,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a6a08091440aaac9bbdb","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-852f97c42d88f1c0044f3365","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"Epoch base and effort spellings for Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7424242424242424","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fcaf3cb464ac261f79da3caf1d671f1dc95f3050","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2727,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=53;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2727,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.24242424242425,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dcb50c57258189566d07","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-33c9238f3ba1df61f8fb94cc","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.928030303030303","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6a486d93eeb4dd25166792cb7888660fe942dd5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2780,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=106;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2780,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":92.8030303030303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c42fc8b653032cc02883","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-03e83b6c267809a4944578c2","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_low","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6a486d93eeb4dd25166792cb7888660fe942dd5","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2729,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=55;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2729,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5cdde891e40eaed129fd","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1863a498d5c035e2ae1d8c37","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_minimal","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8636363636363636","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6a486d93eeb4dd25166792cb7888660fe942dd5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2763,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=89;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2763,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.36363636363636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-793ebd05570ec58d1f0f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-097b31e0df114d9fce5824ae","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9412878787878788","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb3fdf27d42d29ca5876504a02767cd439d256c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2751,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=77;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2751,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.12878787878788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4609d526f5168f253d11","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-42d2e1823ee321552b676b6b","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_low","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8636363636363636","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb3fdf27d42d29ca5876504a02767cd439d256c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2725,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2725,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.36363636363636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f10e70a7abb4c67405f1","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8949cb7681a3e8ac0db9408d","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_minimal","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8585858585858586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb3fdf27d42d29ca5876504a02767cd439d256c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2707,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2707,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.85858585858585,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4319b6e110b20ce7eeb5","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c0154680177237a53b09e0ed","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9482323232323232","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37a4825854a5be6af8702df2e7afe88c56cb4d2a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2678,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2678,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.82323232323232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-22","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-685b4b832963b0894aa6","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b10511e22040e7e61e24902","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14b35d126013743c7cd9baf6106b618baab77230","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2797,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=123;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2797,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-11","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-93813e4bafe4f42f5cba","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77448370ef290092e70f92c9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8781725888324873","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"773623ade5d9a74f770d4fd55f98505d7659a5e9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2793,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=119;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2793,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.81725888324873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3631709f89a7e14d44fa","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-98f3bbbfeb7753d4e81a9b95","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.898989898989899","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ad0c2969184a370599a2a8f6c262b3491e84dfd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2684,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2684,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":89.8989898989899,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cdef938d104027918e28","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-183ae87c492ebef6e3e03ec8","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9185606060606061","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efb5b7a8713a20afff58fa7ccce36c81d3959e46","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2776,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=102;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2776,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.85606060606061,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6bccfacd7dd42b5cde73","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-64c1e22f2613a0aa672ddb1f","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on GLM-5.2.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on GLM-5.2.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_none","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7121212121212122","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efb5b7a8713a20afff58fa7ccce36c81d3959e46","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2683,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2683,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.21212121212122,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-08-14","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5d74cd3df50088c5bb36","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-557a3d96294f8737cef7d950","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.3","modelRef":"glm-5.3_max","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9090909090909091","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"161e3d7c7ca0917060c33438358a3238b3dec753","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2676,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2676,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.9090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ff6609ee1cc07fd78c9f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c84bceb2fe13ad97e717268a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8617424242424242","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa6c7f8ee564167b97a15426452304c9a04f97ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2816,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=142;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2816,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.17424242424242,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5af3d6ce689fa3661aa5","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2f13ffe0f9cd433b1f6ba7ac","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8535353535353535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa6c7f8ee564167b97a15426452304c9a04f97ce","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2826,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=152;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2826,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.35353535353535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-321a7dcb70fd6bfbbfcf","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-976dbb39ac14a38a5127e285","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7171717171717171","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa6c7f8ee564167b97a15426452304c9a04f97ce","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2754,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=80;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2754,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.71717171717171,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f176cd4756f6e9b9c854","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-287a00fdca1dcd34985457c8","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.75","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c53d0de9e31f73ebd60b8543f60988f9f2e6de0e","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2815,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=141;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2815,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a289878034c48e93fd98","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9479d715c79d62d23782ce0e","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7165404040404041","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c53d0de9e31f73ebd60b8543f60988f9f2e6de0e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2825,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=151;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2825,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.65404040404042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5fc221804b37b2d5cfdf","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d8375b54cf47da9e2e76abdd","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7171717171717171","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c53d0de9e31f73ebd60b8543f60988f9f2e6de0e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2718,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=44;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2718,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.71717171717171,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-446e1076c09b73286134","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-84174717ccb9e5ff7e8d40e6","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6944444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5f637894d2261b1a684d5d15dcd9ac41acc135a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2814,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=140;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2814,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":69.44444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8d62ea34bd72077fc3fa","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-077797ea67384b710b20ed94","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_low","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5757575757575758","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5f637894d2261b1a684d5d15dcd9ac41acc135a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2764,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=90;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2764,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.57575757575758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-76980c0e36cfb8afe20c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-defa91fe0ae123728827d062","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6742424242424242","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5f637894d2261b1a684d5d15dcd9ac41acc135a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2827,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=153;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2827,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":67.42424242424242,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fd28f32c36bfb0095bea","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-63f4b9409aa0159db6bfd36c","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_minimal","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.48484848484848486","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5f637894d2261b1a684d5d15dcd9ac41acc135a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2766,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=92;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2766,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.484848484848484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2daea81107cac1542365","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-babf1cd0fc9af0e0b865c09a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8762626262626263","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a67c1b324eb6b0b672dc2cf3ff3d150155836a6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2812,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=138;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2812,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.62626262626263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9f3d7e38729fb53451bf","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3d26d423bf0cae184ffeb62a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_medium","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8503787878787878","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a67c1b324eb6b0b672dc2cf3ff3d150155836a6c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2810,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=136;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2810,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.03787878787878,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6d5a52338e058dec327f","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2fa0d1ff70ffcbeb3169d48f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6666666666666666","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a67c1b324eb6b0b672dc2cf3ff3d150155836a6c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2696,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=22;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2696,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":66.66666666666666,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4744bccb2f1cc36efef9","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-57933c43e351425c4d88286e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8819444444444444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"067e5f4f3b31d104915d10c56080723e3402d0a7","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2803,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=129;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2803,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.19444444444444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1d3b4491c7596f5f1755","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-828e9e8136a517a6f58d1a46","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.827020202020202","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"067e5f4f3b31d104915d10c56080723e3402d0a7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2805,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=131;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2805,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.7020202020202,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2d55190856d56f4fc25c","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d6a007eb77c47a28bb619eb3","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8787878787878788","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"067e5f4f3b31d104915d10c56080723e3402d0a7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2804,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=130;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2804,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.87878787878788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-71b5ab6019aa8eab1742","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0183d18bb4f56e5a2be2fe5a","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_none","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7323232323232324","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"067e5f4f3b31d104915d10c56080723e3402d0a7","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2767,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=93;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2767,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":73.23232323232324,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8a86f814e4bbc7259093","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9a9b027cf977971e1ef6a603","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.914","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"067e5f4f3b31d104915d10c56080723e3402d0a7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2800,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=126;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2800,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2cc080c05a990637ab27","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-149c46f77c9ec1e37e384bb7","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.898989898989899","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e4f2805e6056edd2f53e4fa6ed945c430df711f","selectionRank":5,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2761,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=87;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2761,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":89.8989898989899,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3df9a96e56c63a548519","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5b4fddf3504c0bd2d0e74c00","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_low","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8484848484848485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e4f2805e6056edd2f53e4fa6ed945c430df711f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2762,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=88;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2762,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.84848484848484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-37dcf51c11dbe51414f8","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d0f662a790f2182e67f0158d","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_medium","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8888888888888888","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e4f2805e6056edd2f53e4fa6ed945c430df711f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2759,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=85;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2759,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.88888888888889,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2ef3b37b02138df0c752","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-531f4d0b0ba88a9badcb7357","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_none","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7474747474747475","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e4f2805e6056edd2f53e4fa6ed945c430df711f","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2760,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=86;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2760,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.74747474747475,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":5,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c0b81df241f0da0f1e22","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4dce1779968c1d8eb6a330e2","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.933","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e4f2805e6056edd2f53e4fa6ed945c430df711f","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2790,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=116;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2790,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.30000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-64a20058cfecbc316b5d","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-216d08b9c58b2877cc2e3487","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8358585858585859","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f810bf752383258c62e4f12b8de1a82090f9f33d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2786,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=112;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2786,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.58585858585859,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0ffb84b85a2dc05a6996","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4059d21ee5fa18b0b606eea2","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6414141414141414","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f810bf752383258c62e4f12b8de1a82090f9f33d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2691,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2691,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":64.14141414141415,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7d12c906bad32c34a2b0","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec259e332df90d4223a08fe8","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_xhigh","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8686868686868687","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f810bf752383258c62e4f12b8de1a82090f9f33d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2685,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2685,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.86868686868688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b995b2d5548acccc5fb5","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a1dc6f35831aa78b913e9e77","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7847","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3b39eb4455c879db640915cfd035f67f46dc556","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2787,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=113;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2787,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.47,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-af07510d64aaae640fc7","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2944506b62e65f736039c126","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_low","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7222222222222222","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3b39eb4455c879db640915cfd035f67f46dc556","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2719,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=45;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2719,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":72.22222222222221,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a8e94426db3a1e3d2362","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9e5815eca216d7268485f77c","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_none","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5555555555555556","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3b39eb4455c879db640915cfd035f67f46dc556","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2716,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2716,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.55555555555556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6b543502b32ecdd43294","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b9d1ea64ecf7180e8e63deaa","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.946","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca9bd271ae306758e767851a783d2610596a2dcc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2789,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=115;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2789,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-05","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-99b32cf64441fded3428","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dab8c8a35a0b21a34b015573","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8251262626262627","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8363f19d51f8a23b6973b32dce68dce72b86e5de","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2749,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=75;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2749,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.51262626262627,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-226bc9915ed5a9c71fdc","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-736d5071cfee0c8bb863e192","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_low","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9065656565656566","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8363f19d51f8a23b6973b32dce68dce72b86e5de","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2781,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=107;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2781,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.65656565656566,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-23","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b7c082adb33aa6a6574e","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f5432d156e3b9e881ade1a1c","mappingCandidates":["openai/gpt-5.5@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records append an unknown/none setting to the GPT-5.5 base name; the setting remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5_none","observedAt":"2026-04-23","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7727272727272727","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8363f19d51f8a23b6973b32dce68dce72b86e5de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2687,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=13;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2687,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-04-23","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-49544da3ba2dc6f1dd3f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1f8b2e40563b6fa456d36cad","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8232323232323232","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d6cb8feb93a6f81409d538a62795b8879b1ddd3","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2711,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=37;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2711,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.32323232323232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5cb67dcb909e1af24461","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a73a3be11b5eaa90eb8baf58","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9160353535353535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d6cb8feb93a6f81409d538a62795b8879b1ddd3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2772,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=98;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2772,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.60353535353535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ccc442b0b2b21d3c9553","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a68a92215ded13d15a85bae","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Luna release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6363636363636364","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d6cb8feb93a6f81409d538a62795b8879b1ddd3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2698,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=24;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2698,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":63.63636363636363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-228f98f61dc158de9884","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ad58c47659439a3d8b690ca","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.898989898989899","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa8bbc39d4419880be6c3f64b0d47f6de0d79ce7","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2693,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2693,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":89.8989898989899,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a78987fc2bd154f86463","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4884438b02da1cb2d7c1e622","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9349747474747475","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa8bbc39d4419880be6c3f64b0d47f6de0d79ce7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2773,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=99;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2773,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.49747474747475,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-be776d56f23f3ecb5393","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-440bb0b598aab9d93e79fc60","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Sol release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8282828282828283","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa8bbc39d4419880be6c3f64b0d47f6de0d79ce7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2686,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=12;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2686,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.82828282828282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3d82d243449854d1301b","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b45ca376081c3ec78c83d801","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_low","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8737373737373737","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780a6c5095eff3d48ed56d312e46f77334364e88","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2699,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2699,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.37373737373737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a7f1d08635e1c356c536","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2e28c81b2de675137e7dbb6c","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9330808080808081","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780a6c5095eff3d48ed56d312e46f77334364e88","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2771,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=97;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2771,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.3080808080808,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-50d8eec2a0b1a49817e0","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c43956fc0c1e162c878832e","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingEvidence":{"confidence":"high","note":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","registry":"model_aliases.json"},"mappingNote":"Same GPT-5.6 Terra release with explicit effort/harness presentation.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_none","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7727272727272727","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"780a6c5095eff3d48ed56d312e46f77334364e88","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2690,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=16;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2690,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-08","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7f61b33babe2985f6d3b","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2d4cfc078d2bab08f594c204","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9343434343434344","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b294538785f6cc028c8def623c8042d73c25492","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2774,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=100;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2774,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.43434343434343,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8c29b7ce5ebd295632bd","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a1a617a203f4fb8c4c69b441","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9400252525252525","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f840fd6eb44285b4a0fb34bee3f88e90b295c49","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2681,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2681,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":94.00252525252525,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6ec14ed2f9b150e70407","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9aed01fcdc809785488433af","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9318181818181818","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f840fd6eb44285b4a0fb34bee3f88e90b295c49","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2680,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2680,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.18181818181817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8370ddcfef13a9c5eb1f","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-04bfb0d2efe2cc62c595c9ac","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9078282828282829","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da529a0cb7d34d2eeb174e24ff3362249139d8e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2782,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=108;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2782,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.78282828282829,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-12","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b136fa87cdb687c8394e","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-38b30103af59a4dba627d845","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8787878787878788","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20ebef179090e362130117c0bd29e005b6c4ff52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2697,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=23;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2697,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.87878787878788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f4ff23474cee1301344d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c5e9cfba01527c651e0fc803","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_high","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9191919191919192","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0548bada76d2da2cc90f9e2bb924097ce0dbb079","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2689,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2689,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":91.91919191919192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3ed500435fdec58f17f1","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-39f78d807b6ed31131398a36","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_low","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8484848484848485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0548bada76d2da2cc90f9e2bb924097ce0dbb079","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2688,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=14;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2688,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.84848484848484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8a0f68c24f966e304188","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1be29958869492fd64aff4d6","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9311868686868687","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0548bada76d2da2cc90f9e2bb924097ce0dbb079","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2756,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=82;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2756,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":93.11868686868688,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-34d4a1e986881713f625","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a5a70bfb2cd6e11e556b73a7","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa_diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":23,"rawValue":87.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae0e19eeda2b4922345f8116e1b59e1ef4fe860c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":23,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=23;row=22","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":23,"metadata":{"author":"moonshotai","filename":".eval_results/gpqa_diamond.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":50,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-71c53e994d08162d9835","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/5c/4b/5c4b9bd9-14db-4c11-bf99-9661461459f6.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-6c7296d89a6c8c23da91d62c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":24,"rawValue":87.3737373737,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae0e19eeda2b4922345f8116e1b59e1ef4fe860c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":24,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=24;row=23","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":24,"metadata":{"author":"moonshotai","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":126,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/5c/4b/5c4b9bd9-14db-4c11-bf99-9661461459f6.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.3737373737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-1f4a37852d9a89e96bf8","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-3234f5c09497f3829d541590","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa_diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":7,"rawValue":90.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6dd8708f350ec208147d445b3475f1f507ec0d1f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":7,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=7;row=6","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":7,"metadata":{"author":"moonshotai","filename":".eval_results/gpqa_diamond.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-b4b6cbb0083c13584f70","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K3","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-5a804e9c0fbc819811228413","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"moonshotai/Kimi-K3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":1,"rawValue":93.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b03f22b7880519aeed5d57235a3c1403ba0d0a4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":1,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=1;row=0","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":1,"metadata":{"author":"moonshotai","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":2779931837184,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K3"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4f5f09c9c13ab710ff9e","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-82be1d46353642a44ebaa477","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8585858585858586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"afec2eff5bc9272bbb112fe932ffd373ca3c52f0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2704,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2704,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.85858585858585,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-13","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f986ebfae0a05ff1518b","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a678294d6f8bce35e76eb0a5","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 397B A17B","modelRef":"qwen3.5-397b-a17b_none","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8636363636363636","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"afec2eff5bc9272bbb112fe932ffd373ca3c52f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2705,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2705,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":86.36363636363636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-25","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1408be72c95ffd618e46","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d4506121da75b0ed87b8914c","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8232323232323232","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e3eb1c7cc4c453b448dfdbd8b86173852f40cf95","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2717,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=43;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2717,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.32323232323232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-02-16","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dbfcf05cc2f1268530e0","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9e9fd41e03e75a4c853d11e9","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8484848484848485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"731a6f655f9b70e22dbc7838af32e8f5fa656b99","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2706,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2706,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.84848484848484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-22","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f7b838b7d27f8a4ca3a6","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1e6bfc3d2fe92af603ba8ef","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026-04-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8585858585858586","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55b4d44b6567d543c9a57e2cdef431d03b04ad58","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2702,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2702,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":85.85858585858585,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-22","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3db983bad00fa4804f3b","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a32af5f0fbcabca81f4b43d","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b_none","observedAt":"2026-04-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8484848484848485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55b4d44b6567d543c9a57e2cdef431d03b04ad58","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2703,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=29;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2703,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.84848484848484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-14","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cfad0b8c42217d4c31a0","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bfacb06f72b0d116612b3ad","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8383838383838383","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1ce445bcdb14dd9786dab69952258ccfca7b81fc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2708,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2708,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.83838383838383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-14","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8987be802682487f7eac","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-25735c1c2f2e026a589bfea2","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 35B A3B","modelRef":"qwen3.6-35b-a3b_none","observedAt":"2026-04-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8484848484848485","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1ce445bcdb14dd9786dab69952258ccfca7b81fc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2709,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2709,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":84.84848484848484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-27","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4ccb71194da0566d3a76","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-bce7b2ebd024eeff4ac10f32","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8333333333333334","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc70d30902d41054afc7896443bd6a8d46e75341","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2710,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2710,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.33333333333334,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-616bc5004e0666f290a3","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-55720cbace065dfe3264e9bf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8737373737373737","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"45b75d639ee27fcf3c44c45c08aa96397d014f8e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2692,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2692,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.37373737373737,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-03-31","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3aa9bcacb1ade4a1589d","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6e3812d40a0e6b3886bc49f6","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8838383838383839","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ec39af009a974bf907477f747b6fb2ee3d607f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2700,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=26;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2700,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":88.38383838383838,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-27","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d7ff7ed33854130f92d1","canonicalModelId":"qwen/qwen3.7-flash@2026-07-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-144d4c7cfc6339020cd6e3e6","mappingCandidates":["qwen/qwen3.7-flash@2026-07-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Flash","modelRef":"qwen3.7-flash","observedAt":"2026-07-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8232323232323232","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a91a93f0274bc2db704f451aca522a9ac3c836","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2714,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=40;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2714,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-07-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":82.32323232323232,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-07-27","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6ae5cedc4f5437b51a6f","canonicalModelId":"qwen/qwen3.7-flash@2026-07-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6865bbb19e52f350dd9400ca","mappingCandidates":["qwen/qwen3.7-flash@2026-07-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Flash","modelRef":"qwen3.7-flash_none","observedAt":"2026-07-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8080808080808081","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7a91a93f0274bc2db704f451aca522a9ac3c836","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2715,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2715,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-07-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":80.8080808080808,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f0a68f84abb0e73f89c8","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8cb8550da98d0bc85a39c737","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9090909090909091","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bab96b9b59cb392a9e7779832879a0928d4b195","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2694,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2694,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":90.9090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-02","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-efe71a7bc64d8839d438","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-573447d84bd59bd366bc182a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8787878787878788","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53a497d813ed62bfa0543a64c0aef178d1ec4dc4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2712,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=38;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2712,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":87.87878787878788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-06-02","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2a58028f99bcf8f0c0c6","canonicalModelId":"qwen/qwen3.7-plus@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-08cb6e1981375ff57beb0dc0","mappingCandidates":["qwen/qwen3.7-plus@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Qwen3.7 Plus.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Qwen3.7 Plus.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Plus","modelRef":"qwen3.7-plus_none","observedAt":"2026-06-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8181818181818182","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53a497d813ed62bfa0543a64c0aef178d1ec4dc4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2713,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=39;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2713,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-06-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":81.81818181818183,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":"2026-08-02","benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0b613f1cb976e6b922ef","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3635258e703dd30f599376c8","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.9267676767676768","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f7a7a17c4e1becdd3e4595e26814bbdc1d8c560","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2748,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"gpqa_diamond.csv:row=74;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":2748,"metadata":{"epoch_file":"gpqa_diamond.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":92.67676767676768,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-3e969b831a1259fa522b","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-bf0bc9eb72aecff640badabd","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":37,"rawValue":85.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"236b626bab89c09d0195a3cf4ddd11870e23bd3a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":37,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=37;row=36","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":37,"metadata":{"author":"zai-org","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":358337791296,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-712632480b759df6093e","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/97/fa/97faabc6-a7ca-4cad-8d29-307103287d91.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2e2a183c67063a382fbb1aa0","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7-FP8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":21,"rawValue":87.8787878788,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"236b626bab89c09d0195a3cf4ddd11870e23bd3a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":21,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=21;row=20","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":21,"metadata":{"author":"zai-org","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":358458391872,"pull_request":3,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/97/fa/97faabc6-a7ca-4cad-8d29-307103287d91.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.8787878788,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-58194de5ceca37180c22","canonicalModelId":"zai/glm-4.7-flash@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-ec8d4fb1af24105a3595d65a","mappingCandidates":["zai/glm-4.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7 Flash","modelRef":"zai-org/GLM-4.7-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":79,"rawValue":75.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2888304e553b3f217bd62eb2a22adf733475b09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":79,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=79;row=78","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":79,"metadata":{"author":"zai-org","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":31221488576,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-87dd3fd3956482dbe488","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a89621805e42c57d6e012792","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":34,"rawValue":86,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a17a55314a5a7f71e3646b4d82806c9634c58824","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":34,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=34;row=33","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":34,"metadata":{"author":"zai-org","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-bb308d268fdf6d26ba7a","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/8a/87/8a87d622-4abb-48a7-8ac3-594e31a85525.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-c1d631e567927882fffd486f","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingEvidence":{"confidence":"high","note":"Hugging Face FP8 packaging of the GLM-5 release; quantization is retained in the source model reference.","registry":"model_aliases.json"},"mappingNote":"Hugging Face FP8 packaging of the GLM-5 release; quantization is retained in the source model reference.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5-FP8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":40,"rawValue":85.3535353535,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a17a55314a5a7f71e3646b4d82806c9634c58824","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":40,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=40;row=39","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":40,"metadata":{"author":"zai-org","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":753910024032,"pull_request":10,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/8a/87/8a87d622-4abb-48a7-8ac3-594e31a85525.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.3535353535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-c2770caaccb5a97ec805","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-02f6b891343a5142707be676","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":33,"rawValue":86.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e01c6ab7584e8483c30b231c0aa60549a8bf30a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":33,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=33;row=32","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":33,"metadata":{"author":"zai-org","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-0df1e95d2201afa273dc","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/52/23/522348c0-d242-4517-b97c-4e55c4261882.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a6412a3a306c5d5292a7c9b8","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingEvidence":{"confidence":"high","note":"Hugging Face FP8 packaging of the GLM-5.1 release; quantization is retained in the source model reference.","registry":"model_aliases.json"},"mappingNote":"Hugging Face FP8 packaging of the GLM-5.1 release; quantization is retained in the source model reference.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1-FP8","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa-diamond.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":53,"rawValue":83.3333333333,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e01c6ab7584e8483c30b231c0aa60549a8bf30a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":53,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=53;row=52","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":53,"metadata":{"author":"zai-org","filename":".eval_results/gpqa-diamond.yaml","lower_is_better":false,"num_parameters":753910024032,"pull_request":6,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/52/23/522348c0-d242-4517-b97c-4e55c4261882.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":83.3333333333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"gpqa-diamond","benchmarkName":"GPQA Diamond","benchmarkVersion":null,"benchmarkVersionHint":"gpqa-diamond@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-gpqa-e96094c79c31ac57a4f2","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.2","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2bff2225f45a5bdbff456a24","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"zai-org/GLM-5.2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","protocol":{"filename":".eval_results/gpqa.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":5,"rawValue":91.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bc5062fefea4d4787f608a3ff9e77a3ca2132b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-gpqa/candidates.jsonl","line":5,"payloadSha256":"0e951886a0a40b63e3d2b160988c29c21fa72eab5e447daedfab67c2323a3121","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-gpqa","sourceLabel":"Hugging Face · GPQA leaderboard API","sourceLocator":"rank=5;row=4","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","sourceRow":{"artifact":"hf-gpqa/candidates.jsonl","line":5,"metadata":{"author":"zai-org","filename":".eval_results/gpqa.yaml","lower_is_better":false,"num_parameters":753329940480,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.2"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/Idavidrein/gpqa/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c488a6a177b7d0014868","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e2083acf3133e640402a403f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.6659192825112108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf43b30984302bc9fc19421c58dec3d942e810b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":219,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=36;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":219,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.6659192825112108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-aea72bd1696eaa07a415","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7428cfc243518f3b9e1953a5","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70951d4df1dd541d12abc8281d4a7049967d9854","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1722,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1722,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5f3622831e39e67aec37","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f100a3b9bdf0a3a58aff8ef6","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":16.325411326249803,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fead3b1663173467b3fe42a9c40436ad67621e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":627,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=36;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":627,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":16.325411326249803,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-75f05f6485aa5272a8e9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-108015c827a8bf76f6f03075","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":453.1434977578475,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a06901c727c6c197ab6b0466291c49ec321081a3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1726,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1726,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":453.1434977578475,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-8e9a52b93324d7b6c72f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c89a94617919db33cbe30c20","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":250.73766816143498,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"370d3db5015ab59912777b9c49741d5c68e6a134","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1725,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1725,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":250.73766816143498,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-610d583fb0a91f9f002e","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e057da1c0199c605b046e619","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8aa24f17b5bd51a338f8a8efcb124cae2ddeb87f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1723,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1723,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b247b6871b108649fda1","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-28a4fb57ea854c1b289cdd79","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"16ad374f8f6df73ed0ada582a1dcc79f9bc32c5d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1724,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1724,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b2de10e69b5480cceedd","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-49285b3e172b82c65ce65816","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.7085201793721974,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf43b30984302bc9fc19421c58dec3d942e810b6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":225,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=37;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":225,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7085201793721974,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d591cff6e30a71b42519","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-81d428f5396c2750b51e95cb","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70951d4df1dd541d12abc8281d4a7049967d9854","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1747,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1747,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7edb5ee242ce95ec2a50","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-55f58771a7f4ac232527b17c","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":45.52923426562793,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fead3b1663173467b3fe42a9c40436ad67621e6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":633,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=37;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":633,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":45.52923426562793,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b646e87fa4ac66c75d7b","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-233ba6da46d0adc76c4bdede","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":343.76233183856505,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a06901c727c6c197ab6b0466291c49ec321081a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1751,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1751,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":343.76233183856505,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-24274281fe273eafc270","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ad03caa7ed7616251b512d01","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":272.73766816143495,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"370d3db5015ab59912777b9c49741d5c68e6a134","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1750,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1750,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":272.73766816143495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-efd2f8c24c9cda3fc2de","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1ade467dde068dcb75fee7de","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8aa24f17b5bd51a338f8a8efcb124cae2ddeb87f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1748,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1748,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a51a7689819f759a290b","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-85a959feb68ca147fc5afcc8","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"16ad374f8f6df73ed0ada582a1dcc79f9bc32c5d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1749,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1749,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-31f6ff46082ce8c8bcde","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3375c76b2f233f4d95129788","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.6053811659192825,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b843b9455b4f7d499a4293eb99e018c0a00a7642","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":237,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=39;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":237,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.6053811659192825,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2fb109a2135bdb34e362","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e5031dba02ec96a38620a072","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f3813b977192d82d662b71dbc9182155dee9cbc0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1797,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1797,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-275b504ea86b38ce158c","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b196ebf0d74ac8896fd9a076","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":5.102193982611857,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9673ba6e0074ae7f895379921ef8a2c0e5c8b80","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":645,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=39;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":645,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":5.102193982611857,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-aaa69c439c2ed7294fe7","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b3a5052474884f3abdd6b536","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":524.5246636771301,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f140919177c1b56581ebc5bff5d7899f9ebce04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1801,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1801,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":524.5246636771301,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-805ef38041ac8fc131de","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6c2cc5b09ad56e48873bd81e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":272.73766816143495,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"359af2b1bb2291856ecd7c5fa9b48d150114b05d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1800,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1800,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":272.73766816143495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7e2926378518567030d5","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c271f8a8443a8cd54eb24173","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"986a8885c5939ac55ed9453a6c1fb47ac118ba90","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1798,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1798,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-365f327e62ff335c1c77","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9a427d9e76c479e4b9b3d0c7","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0521e0b72f8b1930389cef0e4ddc711d50856eb7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1799,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1799,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a64e8487b2cca554acb2","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-49aef1f922617d345fee6bd3","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.6860986547085202,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca57a967af2b1e214ba5175892d2656d26d4606e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":231,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=38;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":231,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.6860986547085202,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1e680bf25ed243813551","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f329394ab55db45089725c39","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9379018bc56820996646f61b3889579ef2e7d34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1772,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1772,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-70cc81a71dfa32e351e0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b92e7ba07f20f0c08f5e7518","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":12.414452127318263,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"528ad41dcf29fb286b3252626f139e3557fa6994","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":639,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=38;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":639,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":12.414452127318263,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cd58cda871caeeffbe92","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-82c8b45c20c6b3fc4c7dec53","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":544.2152466367713,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb473f8d6edd9323871401d4653c4853a526feea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1776,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1776,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":544.2152466367713,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f1a9aabeb8493ba8df36","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-65526b84cffa4596621143b3","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":272.73766816143495,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5742722cbde61c64e7900c7e4e5145f6b5930d6a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1775,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1775,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":272.73766816143495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-386a87e578994b2e69c7","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7800a2b48b840cfb073582ec","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9525b3dc571373c90e85b2b2b51a8d7b0222f5a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1773,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1773,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-bfb5a6320de0b935b203","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c47347eae53e0e1c8d613608","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e71408ed8f291074c8192de22d23846b049d4aee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1774,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1774,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-8a2671ad4c56c93dbe2a","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-08d12d4a00361224c17d3ba0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.6659192825112108,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"18b04fdd5285d7812416b02c4e437d56b431afa6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":153,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=25;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":153,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.6659192825112108,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-016744100494e186d707","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-46dbf1518771d86f75e5e2b2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d67e629eb821950b79316e1778f9354a6eebb574","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1447,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1447,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0a4434440b607a5cf344","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d161439fe4613a69e065e194","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":155.438512681311,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a96ad63f4421e200f3cd6d06768fe820eb1d088c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":561,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=25;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":561,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":155.438512681311,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7886ca481f6b9285c786","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-fe76ba2bf1f4445be4ea1933","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d334c782f6630a47acd9452478ff20d5414431e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1451,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1451,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-768935f30f7119d02f89","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d0961d4e356385c6bcaa7b70","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":261.5896860986547,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f03e86055872e81181df467fe24d37e096481873","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1450,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1450,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":261.5896860986547,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b6e248fcddda3148f96c","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ee5451ace7d34b29ff5fae5b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d3fa48c04fed9183fec0809dfb76288a691ee560","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1448,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1448,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a9affe8dca1530f817b4","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3f951470206a8dfb86aa08de","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2af81c314f0e84e19f63d5d2dbb1ba7b6d806889","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1449,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1449,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-61e7a575bbd5dcb4b8eb","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-96eee1d6ecd0b027dff05aa9","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.7914798206278026,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4757338ba3952b482856b16797445586982da399","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":333,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=55;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":333,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7914798206278026,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0a4d86415e16c837ede8","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5827fc592b7009fdac4c9a34","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"437900410f664233bfb31631ce7595e9c15bb580","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2197,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2197,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-46d2a7c5bc87f524b1d7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-16f13900caae2ac83bf7f137","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":57.41822674028542,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e293c82f6880cc7a11e2fa5d2b40dba8bbe14ec6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":741,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=55;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":741,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":57.41822674028542,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-bb22a3ec37f8d7d3f914","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0938a8b2440031aaf2a674ae","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":5.934977578475336,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"68f57c6d6a59fb1defc0517f1a75ee14d6a3048c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2201,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2201,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":5.934977578475336,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a18e2d5101af2e633ef8","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1e47ee2f81b36024014184a2","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":268.15246636771303,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e76807c3d3659db5d39827e9507bfb0da400e71","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2200,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2200,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":268.15246636771303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ff563e987dd86c2e18dc","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-756a5f3bdf51e020fd07b4b9","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f8ae76e4511474cbae1bd0aba8977663cd8d19d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2198,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2198,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-8048248725f605fd22d3","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7daee7e7e08e33b813c1fd06","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f3078d7b04e1c27c719cc9115572a09136b8ee3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2199,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2199,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c2f982cafa99836abf23","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-91b18b348ee385e0fe48fb97","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.7556053811659192,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b1f515629a6b4f9c378927c82d4363d3e2be128e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":339,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=56;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":339,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7556053811659192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9a69b68a3f34d2e373b8","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-87617270986c06eac6c9853b","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a9270c68224afd562e4f6b6e96787cd51ea89271","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2222,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2222,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1cb27dad3f1211a8e21d","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3441a3d70e17fa0d5499ae0a","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":20.737325443280653,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62ab077e1601ffb2d4897349c7ef1b3b43908cec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":747,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=56;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":747,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":20.737325443280653,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-8342f7b8c3c07ac70721","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b168dba2ceb4ab0a86c2897d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":25.378923766816143,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a6e79c6c0e4d3b0d47521f22feacc56c45ba02be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2226,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2226,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":25.378923766816143,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4a8d040057cc333ee2bb","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-05deb1a96744157a1417652a","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":268.15246636771303,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f02d671314cd53ab99c637eb0fa35b2f3b8071bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2225,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2225,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":268.15246636771303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-058a698b481f0af888dd","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-92fdad27cfd634dbb88437ea","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0274ebaaaef7f9a741788ea73f8a463dddae28e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2223,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2223,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3bc4e2296167fb058a5e","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-591fb7af0bac075673eca625","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ade619fb449cf35a5fb700cf105fbc506a58fbe4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2224,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2224,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ab35f25a7317035aab35","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c220f6ba264611824b198ed1","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.679372197309417,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"734b346869ea94189f5be58b9e5d5985a72f4a3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":345,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=57;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":345,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.679372197309417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2ce500e2da57d0bad209","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4ab75c16ca46703180621a53","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"004361cac41fece5a6a88234a829ab4af18a4108","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2247,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2247,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-55dcb393921f7f5a145d","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-bd82dbbdf0b81457f210415d","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":30.2457077674267,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"377f087e4c8436c409c8423e5d6ae169829f4453","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":753,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=57;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":753,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":30.2457077674267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-23238102011edb56d775","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d24f14f769d68fd4caa02e37","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":5.668161434977579,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f73470f18f825edaba528e82275cebd7ef4bdd07","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2251,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2251,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":5.668161434977579,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f714f0e076a51fb5c93d","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4562007455df79ef681da2d7","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":268.15246636771303,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be6a66222516850c2557240370f1d9026895dd54","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2250,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2250,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":268.15246636771303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-94098b998045451f688e","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-084d6e6692f02a5ddd2a6996","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71559923de0f404c8df0ea7f624d779379acc742","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2248,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2248,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cecd01b74e98f698a729","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-415deae519014f5cad8e569a","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2919a30eb5268bfbf40ebc7346736158668c7688","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2249,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2249,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d076bf2e18e47394f549","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1c45812cf388e91df9bb0915","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.44170403587443946,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86119473e32f94ef323af51ea0e9d7b633f54b0a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":351,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=58;column=GPQA - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":351,"metadata":{"direction":"higher","header":"GPQA - COT correct","header_description":"GPQA\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.44170403587443946,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b51bacc9d4f823084023","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-67140daf2514422e91be6b46","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":446.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0051d37ddb83dfcaa1e628b0e9b87d8add3b7eb9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2272,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=GPQA - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2272,"metadata":{"direction":"lower","header":"GPQA - # eval","header_description":"GPQA\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":446.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0e32d0e43e9aa66a4ed0","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ba73b9268d8dd6cfd63c24b1","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1.002433323539426,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eaa45684fc777b9992a64910dff9c5d2ef072344","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":759,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=58;column=GPQA - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":759,"metadata":{"direction":"lower","header":"GPQA - Observed inference time (s)","header_description":"GPQA\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":1.002433323539426,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-fde3c88106cc8cafb38f","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-06d64dd174bc80d7eac00d89","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":5.42152466367713,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41868c537d917c0d55e3b1359d97a8f86a7a09d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2276,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=GPQA - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2276,"metadata":{"direction":"lower","header":"GPQA - # output tokens","header_description":"GPQA\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":5.42152466367713,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-32046449fe7505e9feb4","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3a5af8cde5fd6b8663a4df5d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":268.15246636771303,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8bcd019653959001122c0d7394151cd6eead6c19","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2275,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=GPQA - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2275,"metadata":{"direction":"lower","header":"GPQA - # prompt tokens","header_description":"GPQA\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":268.15246636771303,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-12d1f3f266140a52b0bd","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-18b02e2b1221c1f52d3d7a37","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"607bbfb92a2d061c31b0c94c23711517908b7b10","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2273,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=GPQA - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2273,"metadata":{"direction":"lower","header":"GPQA - # train","header_description":"GPQA\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-gpqa","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5c2f6ea9df95232b2c15","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-52b97e0e0a358c4b9196db88","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a96b102c0c6630eab71ad12197befd768d311a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2274,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=GPQA - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2274,"metadata":{"direction":"higher","header":"GPQA - truncated","header_description":"GPQA\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-bae699b6a70c3120ca5f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9aeb51448163037c677b68f2","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"56db5738f2003cf58c8de5fefbe3569f53ea48c6","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1727,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1727,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-6fb04c9ab0122e9343e4","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7d681990442fe586a7076bee","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.9177449168207027,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"139e4de4badd51d56b98fc99264b923ec86649eb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":220,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=36;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":220,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.9177449168207027,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-af9f3cf666930916184f","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f02e6da48f010d58e989600e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":16.576411149939712,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5639d1ca05a8ead9ab3e1f09abcc327b4e0e5143","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":628,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=36;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":628,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":16.576411149939712,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-1f709d1ceabfcea1312d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7e996bca1b26ac5a68cee407","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":422.7744916820702,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3761eb81839fa81799a723a4e79b6574bb77180a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1731,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1731,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":422.7744916820702,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-313887effa683b3e126c","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-12d61ad4da00cf8165b06a67","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":47.15896487985213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2193cad0684da54800c5afcf3024a64663eefdfe","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1730,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1730,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":47.15896487985213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-9505e0ffa5c3aab08816","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c3697aa048300c6d8441529e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed992ea796ee00070d5f7857276bfb7badf28530","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1728,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1728,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-5eff4d6248477290326a","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-285239fcd6c3d9ca827fea88","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9d7a32be4d589ac13e681e7a69c70fb90a14f5a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1729,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1729,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-3bd0770ae66666468c2d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c8061ce3b588b992be328236","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"56db5738f2003cf58c8de5fefbe3569f53ea48c6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1752,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1752,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-dd9db4f8e2b6f10b19c7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9bbbc8350ddae34f39722927","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.8487369069624158,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"139e4de4badd51d56b98fc99264b923ec86649eb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":226,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=37;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":226,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8487369069624158,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-0823c96736f2f0b2e3dd","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-dd24ca76800ad1cbb67f6774","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":22.45251508421368,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5639d1ca05a8ead9ab3e1f09abcc327b4e0e5143","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":634,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=37;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":634,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":22.45251508421368,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-0acf7756962128602192","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a35d66110c82915fbf180f97","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":403.74491682070243,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3761eb81839fa81799a723a4e79b6574bb77180a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1756,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1756,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":403.74491682070243,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-89d087c1c3570022a8f7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-18d7a44d4b8e457e22ac9db2","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":47.15896487985213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2193cad0684da54800c5afcf3024a64663eefdfe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1755,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1755,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":47.15896487985213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-44bf9f3361d9c13b3641","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0fc4c7a16e767edff5bc6bf4","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed992ea796ee00070d5f7857276bfb7badf28530","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1753,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1753,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-29abf2950ecf4a6bd835","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a41d1b8120b65c93e80c0cb3","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9d7a32be4d589ac13e681e7a69c70fb90a14f5a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1754,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1754,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-718307660923c774da81","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4c568f7cd834f9a264bcaf6e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35b0c89e26bf49f4f1912c16933d071bf1a31950","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1802,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1802,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-18661ed2c0081befe9ed","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a8f6e4d968ddcdf077e12690","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.8012939001848433,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fb78c9f16b2cad104293b3d3683bb562d3bb139","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":238,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=39;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":238,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8012939001848433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-4360e15f3c9895a2ccfa","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ea792de09db76906375e49b0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":4.355410516372229,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c201921f2e76fb30eea8a4def06e3a75e08e1f85","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":646,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=39;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":646,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":4.355410516372229,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-9c7788ac9a1c2a235cf8","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-68c85942a8c524e73833027d","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":390.4158964879852,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"61fb64037d05b9af1e9377f9d4b29d8d8d4cae3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1806,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1806,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":390.4158964879852,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-cc9fb662232feb5298c2","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a042d48bc744986c15b6456d","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":47.15896487985213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"26e6f6a7a008171d5cd59df1b88d14cffc3ec580","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1805,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1805,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":47.15896487985213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-4ee0a43e7cdbd1a1623e","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-81f45c906f303ef4f3080bb5","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5bc388320c94ca2f3461e228d4b453e168f11210","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1803,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1803,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-d0ddc76e3a7c86751242","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ff44ae6aeea29ab99e9583c0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c378e5d32d75b431ef16bd62c6c26d8397457586","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1804,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1804,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-8ac815480b1c5305e424","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1748a54b54d7799e25b983e2","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6540bcedaf62ae68cc17fe784dc543ee89470f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1777,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1777,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-cd8d362b50f2d547000e","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1aba7518f77cc2f9648e43cc","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.8496611213801607,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7955827b3e02723cef4c0dada804a3effa1c951f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":232,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=38;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":232,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8496611213801607,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-691805c2e43066826681","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2931b17f65d2edcffc265e2e","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":10.90394415211986,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"242d30eed97a008e775631335fc809e06ca0c066","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":640,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=38;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":640,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":10.90394415211986,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-5c30af511d8af74504f5","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-99e7b7294ca7f8eff060861a","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":414.63216266173754,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"048ce2a8243d860441195c0a9a7718c239273863","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1781,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1781,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":414.63216266173754,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-be789c4c0d7da78ada7e","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4e21fa39846726355a7d2fd6","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":47.15896487985213,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ce66cf92c8912a8d7d02639734258acfa5659d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1780,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1780,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":47.15896487985213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-a8e6c4b0e7259859ae62","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-df619799122a20091ac2e610","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d4a39f074bc9cce46d1de7bbdb501b9009a65df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1778,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1778,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-8ef2ded4dde5fc7cc82a","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-11074874bb4ad65c8031aae3","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7143f6f17e176b268ead2c44de67378f8d6ff9c5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1779,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1779,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-8b6d738736a568f970f8","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f9cc90e25dd33bc4bc514d4b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81b88dc4b8e87b61ead90bd34abd4f90950d9794","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1452,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1452,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-12466eed9b08a5868e3d","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-56454ddb45e74256015bb93c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":0.7837338262476897,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0363cc364d298c60a233167a0a1d350ec2952981","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":154,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=25;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":154,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7837338262476897,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-da5e01c8cb9395dc39ca","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cf4154f2ee245d43ce393d5f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":33.75197721056489,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"560e1da6d1cf34bf0c28314649dd5961791a0b32","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":562,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=25;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":562,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":33.75197721056489,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-55ec352e6b779f1baf3d","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1a50f1e54e9ff9ceef678381","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e31fe50f64ff274c2c1ba7e212311f8aca4d768e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1456,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1456,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-cf7faf299cfb37f483ca","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2fa5a2d521eda8fdff0d2c32","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":46.208872458410355,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1963a1b1f75fd3e4998afa48a3e217ba889ef5be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1455,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1455,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":46.208872458410355,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-814998c37100b32d607e","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-aa5e6def12186254da0d39f3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adac2b1f7c9f93d0ace6a0851a7855f8d12f2e93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1453,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1453,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-7876e9cdfa0c4c2088ac","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-baee3615440a7b6adb7e5402","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49798ebd55037682fb8d407d1b2d7a1132a258b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1454,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1454,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-0d8b83aaf67eb7c4a08b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5ade6afbfb7c844f821e88aa","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"577e2a14687793daf9a08cab1a88deda4aad5dc1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2202,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2202,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-f2a67146f9c53ba0ec8d","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9d71fcba6cbcc6db457b8e92","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.874922982131855,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df95ced43fbac81bd6d323abff19915f850001db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":334,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=55;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":334,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.874922982131855,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-8e14a8f6f1cb28c6b92f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b964f6dc1092a0d579c1cf57","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":35.937195608664354,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"95e5e641cb59b82ea392008e4aa4ec4112d97f6f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":742,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=55;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":742,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":35.937195608664354,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-f54f2cac57962d2ff994","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8f5887fcc50098747d386016","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":527.6414048059149,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6da394cd8c1cf8f866e462d4c4b1b18243b2282b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2206,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2206,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":527.6414048059149,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-d4fec7662ab31d96a07c","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8139d463558e30db9e6e9138","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":45.67097966728281,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"abd675472d17e81418bb8234d0a6567905c2d202","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2205,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2205,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":45.67097966728281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-823f776507c677cde660","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-62bbcb47ed1b1bfed7b4ea15","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6dc159298b8c826351868a2db87d9187a30324","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2203,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2203,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-c86fc09d34522ec28b32","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c62dc597cc16e4cd0249e00a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcfc347c478f831ecd66f53481c8a8b83518e4d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2204,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2204,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-cf94c9334d9b54d52aa8","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1809a2f5df9d4907b5943505","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d6b6c8e6853b27306bead2049e54a5fa6108f4ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2227,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2227,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-b5c5eedea91c4a8a5478","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-850d09e2057a6f2314c9edd8","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.9269870609981515,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e71b4cfbb26e9d9956568a1c1ac9ce2edeb9240","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":340,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=56;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":340,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.9269870609981515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-78909c0753743874a693","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c400529f2dc9c5a2aeeee358","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":18.83414089833963,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f937505609891411ad90277f5dae3ef717eb5cc4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":748,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=56;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":748,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":18.83414089833963,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-560a1a16cf7c505cef65","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6f98a25a5c1cf68b699bb5d8","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":441.13678373382623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24bf2183be8c18c2e4566b6b157b61b4e8ebbadc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2231,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2231,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":441.13678373382623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-65ef7e8ef4a914eea8f2","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d56ae9a382109738cc5fabaa","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":45.67097966728281,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd3e201feb22f7a491c6d14b0a3756b1533a3838","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2230,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2230,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":45.67097966728281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-d98cc0470a8f563eab71","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ae8b32c0aa016cd1233335db","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b475a262ca572d33e72bff1119439906cd89b25b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2228,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2228,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-8dbccc36f31b720d8008","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-373a165927d45aa72591aff9","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eea8c7ee668d6aba48f81108cb5230b0e47eda09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2229,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2229,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-a94a418852d5fdf123ad","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6a0ca6909c3d103dfa83a0af","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7ccb7f6dffcdb7aa5d4973adde054146f4c27ea5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2252,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2252,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-3bdc3333728ea07a27c4","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-dcc1c565c45f43eb359426c5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.9319162045594579,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35551320098917ae1acfe36883261e093ba02b13","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":346,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=57;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":346,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.9319162045594579,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-03a6dd55d6b0878cd146","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cfe90f169154737c3043292c","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":26.734930773980075,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f3bb6d8d8cbe04d8ed8f4cf8e75a4ad0fd39602","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":754,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=57;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":754,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":26.734930773980075,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-0185c67dc67500ccb097","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2e9063510905cfddbbac2098","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":426.6561922365989,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e790e9bc374106ae74d48b143f2dd766863da3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2256,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2256,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":426.6561922365989,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-04390aaceb8ae260faee","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1859b6b792ecd5d41ee36e84","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":45.67097966728281,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23de3aed4ffea20de6e23de8103ea3b9c86e67c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2255,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2255,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":45.67097966728281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-8f22faac7f73f73b2509","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5177c0585233a841de8872ca","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"36f4b9399cf3d791b4d1b20082331450540223eb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2253,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2253,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-a936bb88809b0568f40f","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-28d114811a2e67f41d11b7ef","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69722e99c87c493cb610ef8fa564eaee9b779248","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2254,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2254,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-00e1ec541d9d0772dcf8","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-88022f74374bdac31b39de57","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":541.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6e3b6f6e645df027fe1e9c90d5b0245219314abd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2277,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=IFEval - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2277,"metadata":{"direction":"lower","header":"IFEval - # eval","header_description":"IFEval\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":541.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-b049c127b61fa77deaf0","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-59561b4573270c470f47025c","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"ifeval-strict-acc","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.9349969192852746,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05dd01326778ea0efad24c748fd178c222dd53e4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":352,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=58;column=IFEval - IFEval Strict Acc","sourceMetricId":"ifeval-strict-acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":352,"metadata":{"direction":"higher","header":"IFEval - IFEval Strict Acc","header_description":"IFEval\n\nIFEval strict accuracy: Fraction of instructions in the instance that are correctly followed.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.9349969192852746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-bb1c23f4485e1dc1bb2d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d36ffa114c552942a44d7620","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":13.15882584436103,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29b649b7db2f9d3787171dd5a607a05f599bb531","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":760,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=58;column=IFEval - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":760,"metadata":{"direction":"lower","header":"IFEval - Observed inference time (s)","header_description":"IFEval\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":13.15882584436103,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-a5572245e91ac4d11bf2","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-44a55b4207de4b3e0ea8bda4","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":647.0628465804067,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e58c8a76f2f0213e4bfa16a89fc751e1eb2907e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2281,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=IFEval - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2281,"metadata":{"direction":"lower","header":"IFEval - # output tokens","header_description":"IFEval\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":647.0628465804067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-04171221ecb26efec267","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2cafe5186cc0858b86956e90","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":45.67097966728281,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"896efc47ee273600c4e56100e85830db297c532e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2280,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=IFEval - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2280,"metadata":{"direction":"lower","header":"IFEval - # prompt tokens","header_description":"IFEval\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":45.67097966728281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-ec99020f77e8b808971b","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-fd56e42363ea92c4625642b2","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40a582c4aa69cf159dd898b6120ec4c8e19f0a94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2278,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=IFEval - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2278,"metadata":{"direction":"lower","header":"IFEval - # train","header_description":"IFEval\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-ifeval","benchmarkName":"IFEval · HELM Capabilities","benchmarkVersion":"v1.15.0","benchmarkVersionHint":"helm-ifeval@v1.15.0","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-helm-capabilities-b7101a1c293cdcccebca","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-bbada9a9b72b5cdba724207d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"109068ab151f5a65ec6a999ba5d6b23f3df35b3b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2279,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=IFEval - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2279,"metadata":{"direction":"higher","header":"IFEval - truncated","header_description":"IFEval\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a558c0afab78de013558","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0744bb54252ae91fc703dcbc","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":20.48127702555515,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c18043c47668f8ae05c1d85c5a9c7323ebd854a0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":625,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=36;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":625,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":20.48127702555515,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1e6493daae7aa459c517","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3b7428148249388f11afb0c0","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.7573661731997173,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60a2e101e961e787907654faa86a16ddb6573c91","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":217,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=36;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":217,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7573661731997173,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b9729f6e6d8048d1eab4","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8af517d3b58ff2f35f07be6c","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":52.297304217949794,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c18043c47668f8ae05c1d85c5a9c7323ebd854a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":631,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=37;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":631,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":52.297304217949794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2f9ff1b52a51995541f1","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4e56d4eced00ec918947be41","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.7800254913409985,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60a2e101e961e787907654faa86a16ddb6573c91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":223,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=37;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":223,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7800254913409985,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2ccc28ccf377a7d94e2e","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8d78d9ab968f70e0293135aa","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":7.381503096938465,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03aa91bfdfc0728bca64d2709c0533528a8754a1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":643,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=39;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":643,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":7.381503096938465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-45b1123aca25e910088d","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9e7332754f47cdb81f431db0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.7167516798874939,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05b3e792cda2912fb99ec152185f046043204901","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":235,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=39;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":235,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7167516798874939,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0dfacf9df043b0fa8437","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a96097a99d87d4667995a83e","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":17.536448448412127,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e3b6d909666ce73b35a6b594c748af0af2bed5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":637,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=38;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":637,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":17.536448448412127,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-10cba75ba1922264871e","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-06d8696d6c867befbfc93566","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.7624408441066268,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b511edf013921c75b9dd2cc7b63784ed30d8c61","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":229,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=38;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":229,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7624408441066268,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-df51063febc2e528ebf9","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2c94a114023f7006664529ce","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":115.28182297150872,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a15c376096355d6acb602ec2a336763798644f94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":559,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=25;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":559,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":115.28182297150872,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f0d7a3bfc9f344e949be","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a163c1ce249892d2865b14bf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.6989491402703004,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"658e72b02a9e4114c7144e2f2f7edf75a80a219f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":151,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=25;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":151,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.6989491402703004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-dbbd8b45fa04748b4da6","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c1dc9926b50781ce4fdf76c1","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":74.66990821942755,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"602485e7dc45d24cd268940efb8a41b59cbf20d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":739,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=55;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":739,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":74.66990821942755,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f0887d11d1d19e537080","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-24028321c200b6a2eb5e105c","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.806750930922304,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b75557e2b0cf1180c2bb37305b7d58fcdd24595d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":331,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=55;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":331,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.806750930922304,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5c06dfc3301045087173","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a032dcce84d68a02c8069cce","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":28.206869066978612,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90c2e2bc0347697d4645427b403410e64c25c0c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":745,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=56;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":745,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":28.206869066978612,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b03710a25825b0187807","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7448497435490e16f84e6134","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.8189777476920754,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d26cf7c79fff887e455d034e7e63763b260864d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":337,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=56;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":337,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8189777476920754,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-df7bbef8eac1aa9f9ca1","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b71b9350508e753a88be92bb","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":48.213836350621065,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de3b0402100fbad7664fcfe46bac2c535ce137f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":751,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=57;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":751,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":48.213836350621065,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-374a3c56132c4e754384","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e9864a3857d42462be85fddd","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.7483391618552583,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7b798397a751ee30b1b48f1f6652461fd28c300","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":343,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=57;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":343,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7483391618552583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-96386fa0a6aa67f3fb9d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a607de0b4e8dada34313d893","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"efficiency-score","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":10.620566227529599,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"164cb99f9a536c8ba3495c9a71fa02a078258d0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":757,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=58;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":757,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":10.620566227529599,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mean-score","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-884a9bddb818eb8efb21","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d20e3d2e031ffcf1cfb9883a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.656429079920834,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"67951ccb4fc2fdf6387c922542baa63cf2720896","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":349,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=58;column=Mean score","sourceMetricId":"score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":349,"metadata":{"direction":"higher","header":"Mean score","header_description":"The mean of the scores from all columns.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.656429079920834,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-271816c4df4590d8a974","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4dc1b4cde28b8e8d68fb2bc7","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.859,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5dcfeb4297e933cd23a19a63b1a787f28fa0efa5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":218,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=36;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":218,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.859,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0f3bd4a48bb4f62d1ebe","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-62463aece6b6a11e06306368","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05443ddb7ea6c2b01e524023036e7bc4328cf5ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1717,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1717,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c30f8a0dc6665ed54db7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-dec0ae012fef20af278256bf","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":12.630421590518665,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb023b44aab3356e744636bf46526f6c2a1aa005","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":626,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=36;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":626,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":12.630421590518665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-676b7135622274b9c616","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a2e6aaf608e6af7848d91dfb","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":344.469,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4cb692e0958cb6eab6dc7a7bce9e831170aa925","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1721,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1721,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":344.469,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-57c5ceafef9b7f0805f3","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d0cb144d91b5cca2f3b781f5","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":230.461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7bab4a685c61784705dad42ac10021c29589aecb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1720,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1720,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":230.461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7e61c06708283cefb748","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-323c532b00bee63083110874","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c1b578e6816d3838059572a9a07bdba71a084c0b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1718,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1718,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3687e63513f6548cb981","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8a3dd521fa93b0156a10e0fe","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4ab3b6c7be703d49d3e384afbc9a562415deefb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1719,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1719,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c00da061fc745ad2cee2","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-29b092e9ce38c30f03a0a717","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.875,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5dcfeb4297e933cd23a19a63b1a787f28fa0efa5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":224,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=37;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":224,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-73bc1dea50b8b8d527bb","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5bb38929fde4df70d9181d85","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05443ddb7ea6c2b01e524023036e7bc4328cf5ed","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1742,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1742,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-372cefd2bf7729f403d9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b78864ce6a16548874675ac7","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":28.46593898815197,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb023b44aab3356e744636bf46526f6c2a1aa005","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":632,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=37;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":632,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":28.46593898815197,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1b675d1454065ef050df","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-75afd06b104964828da09508","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":272.871,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4cb692e0958cb6eab6dc7a7bce9e831170aa925","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1746,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1746,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":272.871,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-037c52c18a70e61211c9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3c0899e5401552537ded1fe8","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":252.461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7bab4a685c61784705dad42ac10021c29589aecb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1745,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1745,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":252.461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7db4b0d0bc9f27d9d78d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5aab4f6497d484e2120aab12","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c1b578e6816d3838059572a9a07bdba71a084c0b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1743,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1743,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2bae282caf69ed2efe36","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-655d113fdcf0e5bd6b52c7c3","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4ab3b6c7be703d49d3e384afbc9a562415deefb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1744,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1744,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-e61432d9e47846ecf151","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0cce23b25e106f53793294a9","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.777,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b21c77d84135169e441a663f357f06d2ab005da0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":236,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=39;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":236,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.777,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ae1526266b78736d2d73","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b705b87ee3fce8ed31510038","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e03a8376b3711cc26d21afaff3f1352d5de5de7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1792,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1792,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-6e69ca74ce39d6c19ae8","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b2d87261d27f3b44efbbc748","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":3.7008020806312563,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed8b29d1b4f645a273ad38702c0f290855383e73","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":644,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=39;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":644,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":3.7008020806312563,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-79579d50e03063d488b9","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-944b0b5de2efef1acdddccae","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":374.129,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cabb40ad90e302c2ab62d0e609578327a43184b9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1796,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1796,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":374.129,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-6d6b397f800e268495f4","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-964552be6d1b189d0f1ab2cb","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":252.461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ddb70e8b73546cd811dc7d4031d4c996668ef9f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1795,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1795,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":252.461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-6999906945a195391b75","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a6a1986c54ef3f0e565bea38","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba53bb5679848d94f5eea9b37f8db811f3cdc28e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1793,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1793,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b33c9008e3409a54c325","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c32e7b1257d82f1905b1bf5b","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd7cf4a46a48089cc2ca5a12196975651cba526c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1794,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1794,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d771df7b1f3d491ec505","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-28b85b02e2a048b002fe7473","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.869,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a0b071150ef8ed033f70d7be379eab6f384129e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":230,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=38;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":230,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.869,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-6c506a26bbad02d996e4","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6ef956645a616556cd151b4c","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4d57e80edb313871fc6bd18ae9d1c6c0c35ffa6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1767,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1767,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cf53242ed87895181ce5","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a583d4a278730509df2036cf","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":9.029817205530268,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8c286951bfa40c156ffae095ccac84aeab8a010","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":638,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=38;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":638,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":9.029817205530268,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cd90c95e1b51831637d7","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-bad6b08dba68336dc56dab47","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":392.292,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5421388d7a07576fc0c385c5e365c1453cdbca80","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1771,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1771,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":392.292,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0e1328c9f8624f64f119","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3b8b4976704354b429043442","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":252.461,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27d56338ec24683ebd9f8643dd93d7b0778440c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1770,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1770,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":252.461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-6917cba3c61d2bc5bb9c","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cee85663dfe60953484b028e","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e53585ba6d46b36d822ead8c915b575eff800089","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1768,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1768,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5269b1d6a7f18bf50356","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0ca0a247695a3a1d8140bdc1","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99dd0dd3938d9e19f4dbe24534421fa793f433df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1769,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1769,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a0e38d0ae714c6fb3101","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-33666e1c9f1f868f19a8cf2f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.793,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05f7f54580bdbcfafbf1225b2d10e528542cc155","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":152,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=25;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":152,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.793,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1dfa8beac8ecb93e6534","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a0bca293f040055e8d1dbf08","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fe802b0e9b7f109b87a44ddf99d84b8c1501fc5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1442,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1442,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2b857cba2106a1a6c065","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-82de2f2745d727cd8b849237","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":91.01470815229416,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f67b457e632b0362ee4f6288f0d545ce94df059","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":560,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=25;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":560,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":91.01470815229416,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3b668f42a98986238438","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-363e3c1d2807a8f55f6de709","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55adeaac3794e8256c2b74175e6ec3102e21c3aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1446,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1446,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-65a022d4f0a62040f7f0","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cddf08deafd23875e988cf10","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":248.757,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"981fb86019b2f5af16e980a3cc8ba4a4ee91c871","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1445,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1445,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":248.757,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c384ccc89cba975bf224","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8c9c8f49485895fb51e7630a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b6289e122cba26837bafdf47ae96b2e81894c877","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1443,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1443,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-95d2d6fbc20fada83989","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-148d22fe03b032d2beaf3be7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63dad17e72a22be9317907a5b1f33f5c21694af4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1444,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1444,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b528ab915b096dd513de","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-10e62447707d9f3df009a411","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.863,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f15c4dc853d2080e1aa13e50436f3729d1538a27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":332,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=55;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":332,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.863,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-bde95cb5153bee04e06b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d2aa47b70002312d379b1b46","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87037306de6ca8c24cc9558a6a6c6411e831b68b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2192,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2192,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3b273c51ff7767506a68","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9c4568356dba7a5b1f335428","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":18.668269051074983,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1e8ac72bbe2a2b208a720a91e46236445f15e345","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":740,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=55;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":740,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":18.668269051074983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c1a77d2650d11530aded","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-772456a5e37d5e4c4d2191c4","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":5.028,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d7906ba8b0f4ce7b1f2d8537db302faef9e49eb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2196,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2196,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":5.028,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2ee1f5a1476d1379b038","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-de75be1bce3657c2fbbea246","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":248.569,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e2c1144e96164fcf9f112f2ac565ad4373ce1f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2195,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2195,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":248.569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5f4e5f93d00ce6ec105c","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c6c3e6817bf9c8428d52286d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60d0769f2772402fbaedc494a8226ea17313e238","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2193,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2193,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-bdaa296c247a9c2e7b5e","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9050ced713f303c28ab09883","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3a7f12280eed74fc5042e064e9124dad8b1a224","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2194,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2194,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3d54d972d1ea166050c4","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-799ea9bda61c664b9f91a0dd","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.835,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5526fb04326fc423bc89f17d2988f46fa9ccfccd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":338,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=56;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":338,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.835,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f1fa731c53e9656ded18","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e428e6903c580502cceffcfb","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1735827ad923876969a7653a23da78052b782832","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2217,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2217,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-66e89accda83cd422c30","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-afceb5ab1850eb4c52976bf0","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":11.802515007257462,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"31de4a3545c1d41df92424b79ebbfd810649ca90","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":746,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=56;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":746,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":11.802515007257462,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1a7fd417aa0e8650b480","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1845f4467c1af6c2a692d213","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":17.495,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2aa568ce53cf9644fb67a693a3cd619bce1b2527","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2221,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2221,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":17.495,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cdbbc163e2c74002661f","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e06c90d4482364cf13ff8b06","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":248.569,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ea60be516f758c60947b601a4e8c15d7371ba5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2220,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2220,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":248.569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-70eba4821615f978680a","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-924caa79923cb31fe5bae007","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e4e418a73c8e335aca7d35660e631f910ed400b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2218,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2218,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9e544306a13c0a7082bc","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7a368fab015bbf6490381e20","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"041735d1232876cc4381b132d78c29df7a1a6f2f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2219,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2219,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-390bbff67d42ff9ef6c1","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-699bd1d9ce3dfee555ea2566","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.778,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49c49b640d3c53b26344cf928feaed96a2c24521","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":344,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=57;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":344,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.778,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-6484396b5efb4d8d30f8","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7427dad156fda777a7a96427","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c214a03d6b0bde586f0d37eeb7b67d908df2bc9e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2242,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2242,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-31af202cbfadd3d88911","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7355f22bbcec5cf659aad4a6","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":17.336622306585312,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2d737321b9b3620bf1fe7c85eab2a98743c748c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":752,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=57;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":752,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":17.336622306585312,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-07b97ec826447a82bdda","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-86ad5496365142140e649e09","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":5.385,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8d23bd4b64ece25c35c3d0fd429b88b77833e165","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2246,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2246,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":5.385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f26f913ab6f30d491a27","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-eaec22cdfdd9b7bbdec893c5","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":248.569,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"75f9302f98c7a386e61728d61500f15f73c61739","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2245,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2245,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":248.569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4ec5a1de4cdd81d89793","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-fe43ee796f849b5f5a37bcc7","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35af4f9188b7fbdafb7108566837dd4bf764b23d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2243,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2243,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d70026d74ec17f27a2b8","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ec043a7f3eea1991e31d2fdc","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e7be6158d8b49b1fc3a36447b15d481d2a24d6b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2244,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2244,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ddff5905b71e6f21aa25","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-aad1b27d6869cf445d4c0689","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"cot-correct","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.579,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49fab80d2842141713ccf407ac77b9fc25c109ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":350,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=58;column=MMLU-Pro - COT correct","sourceMetricId":"cot-correct","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":350,"metadata":{"direction":"higher","header":"MMLU-Pro - COT correct","header_description":"MMLU-Pro\n\nCOT correct: Fraction of correct answers after chain of thought","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.579,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-86903355b99726e2806a","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d4700a3b45db4641e1a326b5","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e1b14e08167da9c55da0b18b3f6f7776ef07ad1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2267,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=MMLU-Pro - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2267,"metadata":{"direction":"lower","header":"MMLU-Pro - # eval","header_description":"MMLU-Pro\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-dc05b1ddc9a567fc7d9e","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-99390f292e52c2eadf1a4cbf","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1.1470122172832489,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8877fb86f5171c8cb628a0c4c208ee34013cad16","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":758,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=58;column=MMLU-Pro - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":758,"metadata":{"direction":"lower","header":"MMLU-Pro - Observed inference time (s)","header_description":"MMLU-Pro\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":1.1470122172832489,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1c097230803c7736148a","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a16516a3fea25694e2b5b038","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":5.002,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88f639edd4526a6753229c7c55aabec453787ce8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2271,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=MMLU-Pro - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2271,"metadata":{"direction":"lower","header":"MMLU-Pro - # output tokens","header_description":"MMLU-Pro\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":5.002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a088c5a30770683a2db6","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8dd7c1139fc959d8cb88289f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":248.569,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"500712b2d153b5c48fc6373a4daae433e3fb3261","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2270,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=MMLU-Pro - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2270,"metadata":{"direction":"lower","header":"MMLU-Pro - # prompt tokens","header_description":"MMLU-Pro\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":248.569,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-716c77c864df0446c2ef","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b7619b934c250742f7a9e84c","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b33e4841ccbbc140720ad5d99f6d1b3dde6632ee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2268,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=MMLU-Pro - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2268,"metadata":{"direction":"lower","header":"MMLU-Pro - # train","header_description":"MMLU-Pro\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-mmlu-pro","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-16c79329ab0837cce634","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6cb5c1a4e608d9202ec3d10b","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72782212b042b5ffa566157faadd2ce47626454b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2269,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=MMLU-Pro - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2269,"metadata":{"direction":"higher","header":"MMLU-Pro - truncated","header_description":"MMLU-Pro\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-08202c0746ec027b3e79","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-31662e4d900d8c1a52e33ffa","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.5111666666666665,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7088caf61b5f6035c0a55120720bffd953ac8dab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":222,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=36;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":222,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.5111666666666665,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-85b0547315f4cac09ed4","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-66b83c8b54e04588161bfc6f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb18848131c82240ef9440ba4b6ccda4ffd5bd72","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1737,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1737,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-29252ccea9d7463c0fe5","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-10505c08c9f43b98544806ab","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":27.025822179079057,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"978ccd6f7475c0e09febdc47f6983252966e27b7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":630,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=36;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":630,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":27.025822179079057,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1e3c4a19593e5e79a482","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-48d5c460ee7dac4d52c0ac5e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":893.894,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7d7af5cf6b1088655c76ca305e7e502351739b5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1741,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1741,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":893.894,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-20ec4ed38df5e0675414","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-21dd46890ff147e8e6a16bba","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":110.563,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86f4644376734f84282e53fad9666112afb8a684","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1740,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1740,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":110.563,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b0b1169f9c066922be03","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c441237dd23706e9e2bb5954","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37e8cb6614c07886a79a3b1cadbc2e49c2003a89","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1738,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1738,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d0f9addffc219e897c14","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-117976cb7e525365b7b10e1f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b8850bca59f9c9773ca8e00df2156338f282903","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1739,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1739,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-897d3b95dea64801bafe","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-69cdcf08bd40ebb7dc598af3","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.6160000000000001,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7088caf61b5f6035c0a55120720bffd953ac8dab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":228,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=37;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":228,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.6160000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-112fcb8da42dd88e59b3","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-77504be8cdfaadbfc72bb5c0","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb18848131c82240ef9440ba4b6ccda4ffd5bd72","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1762,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1762,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d3b6438d132a5aaa0256","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cd7119a265dc76226677de61","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":114.84836924410313,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"978ccd6f7475c0e09febdc47f6983252966e27b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":636,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=37;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":636,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":114.84836924410313,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7541e34036d4f7f3bef5","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-24d0e101d3b4e9265d45ea93","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":691.066,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7d7af5cf6b1088655c76ca305e7e502351739b5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1766,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1766,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":691.066,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ca02140f589798cecbe7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-72027dafcd1b2b77db6fccf1","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":110.563,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86f4644376734f84282e53fad9666112afb8a684","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1765,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1765,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":110.563,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-77a2fa7447f7aa5d2374","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2ea4f1df14cf638913dea37d","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37e8cb6614c07886a79a3b1cadbc2e49c2003a89","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1763,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1763,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-82e0854cde93b6bbd2de","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9ef46f0e16eec205e4a59ae5","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b8850bca59f9c9773ca8e00df2156338f282903","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1764,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1764,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-857c0bad412390aa98f4","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1561a5303812a8bf11ebac16","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.561,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df4ea122d3cb44abf1bdbf150745b006fa198781","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":240,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=39;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":240,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.561,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9e71c1b060b16b42ff3d","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-875ba97c748e3d9f4d8e5eb3","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"12cbae177b94a6ca950be33aeb38a3004689263d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1812,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1812,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-dd76ff9fecd2d8679a83","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a70f394d7910fc54ca8be1ff","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":7.431977860689163,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8fd4a2d5933b216a42bbb9682e03ba806b41e907","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":648,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=39;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":648,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":7.431977860689163,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5c80b1ade84e6d63fda1","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5408e32576fa045a8f11758e","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":937.799,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa416904d2d75c09922c673afca2c81643b10182","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1816,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1816,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":937.799,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ff178bda1f974ac17f73","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-422c066cce08a665dbd0aa3c","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":110.563,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32c16f0089fe5d19dd21d97cd27a544ba31962c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1815,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1815,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":110.563,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-331c0089421d125448b9","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cb292d1e255b7b4bb303d70f","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8f866243d078e2a3ccbf201725390157bfbeb8f7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1813,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1813,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ad9294c62b46ec33f9c9","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-346e1efd2a30bcec8bd7d892","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53b73268fd50da788e4cff6de11af6b3fa2a8731","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1814,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1814,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-52d020650b5a0c389409","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d25fedfc3955d1313d1de2d0","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.5529999999999999,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb1361039f5d55193ee2e835332d197b6e13b11c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":234,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=38;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":234,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.5529999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-276c520f07348f7fdd66","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-519ed1d141faa9c375e8a095","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b955edad4d2c0a41478a85290df09405efccc9f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1787,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1787,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3fd5ab0ac28f6f9117f0","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-69d00d4d8792b8298c611023","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":16.790386716127397,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72acbfa0d7cdd306515c5e8ad43d3ffa9f50a6f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":642,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=38;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":642,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":16.790386716127397,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7538e19d7c7240a8abb1","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3fa658a3332268b9eb35747b","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":892.774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f0aaa9c72356f5b36969ab8a14b5d356d87df4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1791,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1791,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":892.774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2a4205f173ee5c58c030","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-183d0c508dd9a998deb00939","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":110.563,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc7581ef2012491eb04431251a04ab2323fdc93b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1790,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1790,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":110.563,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c4b0664ae6659f0c0609","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2436f0769bd2883f88f9f966","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76618229bb87530c307cd33bdaeabb69b8f8bc45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1788,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1788,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-64d4ece6bd07d64f2481","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d08d9ef2f8e1d8b598d83f51","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83cd46f3cd7e7bb5b9e9b196c361b386e4158a47","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1789,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1789,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c8923aa08b99ad272abb","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-13ecfc101c0c14fa346f1360","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.42400000000000004,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fe1c6487f17fce08930cbf5b7875b4207fa9bacf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":156,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=25;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":156,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.42400000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0fce976a0841d6feeb8a","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-414846420fa1d41de7d49b7c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1506433b5db5673bc9d6bf0405041a4238e5504f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1462,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1462,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-940d10ac38875d059f19","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d2f005a1c9b325e9429522d5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":208.35547973060608,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bc6969b643ded19a20db3edb7e5d1bae91970e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":564,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=25;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":564,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":208.35547973060608,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d60aae3601d4827c689a","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9b893874dce851cc5a381a63","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af6c9479335f22ae06686f6f4b5394324222874c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1466,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1466,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f01d7084526b067ad43e","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c1faf3f22855bf4ae684b8d9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":107.102,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14cfd7a278126833cf4651c3a181ff39e97fd6bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1465,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1465,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":107.102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2ec1292654bd919410ea","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cba52e1facc7e7079c53a1f2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab4a72c4b05e1e0d16328c2bbb7bde4c91be6b73","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1463,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1463,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2ccae547c11eb6e34566","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-06c98d8815a52459b90db745","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facddaef9dfb5b206f743e900c4c1289f3a4ac8e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1464,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1464,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-954db20b668249ef3d6f","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c2bd026fd547d3b2e9832fc9","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.647,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"add4819e068715f4207ef454e89ebf298d3de3c5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":336,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=55;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":336,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.647,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cfe811627a1a2dce2d9b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-86072d43c0dace2ba8c9091b","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"edaf6651bcbe32a8b7226456b75367c866c9eba7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2212,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2212,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-53347e67a64e7fd3693b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-74737fa1b0410fc97cd61ca4","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":172.73094402194022,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"96cbaae93028ddf79b70f5af784a6bd2dc817d8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":744,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=55;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":744,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":172.73094402194022,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-304a399430161894e689","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2e2198a8496a8ad40eab5afe","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":228.774,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1836a822e806746bca1c1acd5282178eeb5a2542","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2216,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2216,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":228.774,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9bf4b1e7f6544afbefe7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9423242d972a94d57ff043de","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":109.623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de83a989bdacb84a8af47d7c5e6634f1fa9a7ea1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2215,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2215,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":109.623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f244f20d8b3d343eb489","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f30bf51fa79d3d77b5ee4cc1","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86959b3626d46eb868bd2ec1ea357b81f1ac3086","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2213,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2213,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7d905e853baeb39fd988","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-45aca89a2eefbbabfebcd5ac","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f5df6f83f7791395868c1917120c9c743274746","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2214,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2214,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-8eaaa027206363ae629b","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1107fd367e7c964aae4f7417","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.7220000000000001,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6610137495d8c163c76c233f305a8e2e57dc08f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":342,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=56;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":342,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.7220000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4da36a3d0b88b9b1d293","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1cbcd2e4a051908e985935a7","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dbb7809fa836ea4822fa33106d2bed4a7148257f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2237,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2237,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ff7fe06f1ecfa3a4d46f","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9e39d39ca2e640ae5a59185d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":55.806421615123746,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ebee102890a1411aa491a1fa34e60addc92cdb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":750,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=56;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":750,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":55.806421615123746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-76fdf3fda2117ff5f3e0","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f371f5236e465873ae63d91c","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":362.654,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4c2b393e600c4d284363593784c70caa32530cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2241,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2241,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":362.654,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-86f3d64482ca6ebf9dee","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a41284a985a12894b76ac610","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":109.623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bd86e02e357c2602521691deb17f6371bfb341d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2240,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2240,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":109.623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9848109bd299fa5368fc","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b2bc9d4217a8072b7a36e131","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20d06a110f3ae7446ffffff16723ea46543662d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2238,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2238,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-670d78f92c6978681d7d","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-903f2ce4fba5797237a2d119","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4780aad6d1f65493c6eb2984955657ba44cbda2f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2239,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2239,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1179dfbb61444eff4ad9","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-735c2c396def5db2b87d3a15","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.5465000000000001,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63a7990b3c8e9cc92adca50404445a9bd3ebb31d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":348,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=57;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":348,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.5465000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-b94c3ef3da7c3fec7d78","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2dfad70de6362491e8e0277e","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00670ecd2e279c92c2ca8146159448e7f257459f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2262,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2262,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a2a7a6750f1fe8beb8ca","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3bec7de6fa401b29f519bc0f","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":119.19145288252831,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bd4fd78ed3adba8235b287c5cfce8770bca142fc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":756,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=57;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":756,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":119.19145288252831,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-01f770ff31d415bc2cc8","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2d5f35b5b2592cbcb8037cd0","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":222.15,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60621a5cde15a9bd6a6b1f9f46f93cc81b7a77fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2266,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2266,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":222.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-eb873efce48c548e0a10","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e42e0eaec3dcdefd00bbe931","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":109.623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03cc2055ad5f951fbd1d8bdb37aa09e609ba15c4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2265,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2265,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":109.623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-761696b57f11ce7f4dc5","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0304d441e12602650eb7fa62","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eda51f9028f3fc4e2cabc11928b2e9a06bc3e1a6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2263,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2263,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-e1830c6b883ededbb784","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-22183192965a46c0a1993e4c","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"24e241516ffdf68f635706d460c36b3d1f254fbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2264,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2264,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f7d9674203cab6888124","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4ec707e7a685f5e56a9160df","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"acc","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.4635,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d2cb6ee921e74716643dfd3d4b7eb58bc25afbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":354,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=58;column=Omni-MATH - Acc","sourceMetricId":"acc","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":354,"metadata":{"direction":"higher","header":"Omni-MATH - Acc","header_description":"Omni-MATH\n\nOmni-MATH Accuracy: Accuracy of the AI output judged by GPT-4.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.4635,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-32cda2ff9adf958aaf26","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-967b97b7819c680b538c61e6","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea8f806c114ee7ab51063b9a394a7db39d7814f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2287,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=Omni-MATH - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2287,"metadata":{"direction":"lower","header":"Omni-MATH - # eval","header_description":"Omni-MATH\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3a1c34033a4276861b99","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4c64b2829d7b757bee94c930","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":9.713221177577973,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d695899acfd4cf46f435afd59c44728af4dec73","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":762,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=58;column=Omni-MATH - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":762,"metadata":{"direction":"lower","header":"Omni-MATH - Observed inference time (s)","header_description":"Omni-MATH\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":9.713221177577973,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4500d63f0070c1ac97ed","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8dab701e7c16bc03aeeacc53","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1256.266,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"64b968f43bad13172cf811c5f269a458ddb4f1cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2291,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=Omni-MATH - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2291,"metadata":{"direction":"lower","header":"Omni-MATH - # output tokens","header_description":"Omni-MATH\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1256.266,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-e31332071849e633db7d","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3b1dc0aabcee6333f6cda88b","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":109.623,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"19e5f82d3556f2e526669864f6c5a386cfaffb33","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2290,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=Omni-MATH - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2290,"metadata":{"direction":"lower","header":"Omni-MATH - # prompt tokens","header_description":"Omni-MATH\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":109.623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-512be217f28624414aa8","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-10886e90a22f028b3375b504","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7ebbc84f746d988dcbacbb9248e510779d261736","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2288,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=Omni-MATH - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2288,"metadata":{"direction":"lower","header":"Omni-MATH - # train","header_description":"Omni-MATH\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-omni-math","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7fbb6280c611e7c98b02","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-eaed9d1e52fa6cbbeb517c7a","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4ed7157682c65bcc7d1c4cf0ff778704e85b61c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2289,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=Omni-MATH - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2289,"metadata":{"direction":"higher","header":"Omni-MATH - truncated","header_description":"Omni-MATH\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c0bb939f820ae208ccf8","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c78f0fc943e44b907a821343","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db79aa1b169cc26b2d5a80f68a0cab50703c6c55","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1732,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1732,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4561ed0c23317587bdff","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3ab34fc5554f75b5ae31698f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":29.848318881988526,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9b8a9f6b09935c3308a79e9ce7d683321f9cea2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":629,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=36;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":629,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":29.848318881988526,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-969e2c0179aebb0b2791","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2994752cf2d410c24d71acd0","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":936.927,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd40451f19842bdd5367e7a692fe9bf1c9a3ae34","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1736,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1736,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":936.927,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c8344ff3984ca4011468","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6df75c17d8a60c9865882b70","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e39b2050fb2959373e9b0a5906458c36ff0adeeb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1735,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1735,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2944fa274df8e3426e7d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0cb9ad9b268d8315e1fbaa8f","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88628a7cd05d6103ef11247598ffe2032746013f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1733,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1733,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-516ec89422638a9062fb","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3837237220087265149c4371","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"651641c542317a195bb7ee47a0df1150c119ef3e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1734,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=36;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1734,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-593c52ee8382977729ff","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-41aaa3bc3e91b6399ccdea3e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.833000000000007,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0afdbe5409be4f81dd76c30daf01f031cfcbb7d5","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":221,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=36;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":221,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.833000000000007,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d46aa8361a32593400c9","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-fe23a2c6ce43111560fdbb39","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"db79aa1b169cc26b2d5a80f68a0cab50703c6c55","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1757,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1757,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1ffa1230558d7e5a703e","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8029c5a8992e11f81538f203","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":50.19046350765228,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f9b8a9f6b09935c3308a79e9ce7d683321f9cea2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":635,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=37;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":635,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":50.19046350765228,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a3c600e3ccb0d21877fa","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-cdb369a3a411a40135285d19","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1195.769,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd40451f19842bdd5367e7a692fe9bf1c9a3ae34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1761,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1761,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1195.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-df72f1b8e7e490d3ea70","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-416a5ad67aa7b0d977e60523","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e39b2050fb2959373e9b0a5906458c36ff0adeeb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1760,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1760,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-448db06530eac25e9cb7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-bb6b1095638070ebaf2be889","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88628a7cd05d6103ef11247598ffe2032746013f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1758,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1758,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4e2f687b5513df2f6ffa","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-7c9708b75337ee512fec0e9e","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"651641c542317a195bb7ee47a0df1150c119ef3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1759,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=37;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1759,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1b446b114e8fa8302aed","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-fe794580cd60f4cbb378edcb","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"HELM labels for the dated Claude Opus 4 release; extended-thinking configuration remains visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus (20250514, extended thinking)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.8518703703703797,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0afdbe5409be4f81dd76c30daf01f031cfcbb7d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":227,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=37;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":227,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8518703703703797,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c2f9bc864cb377e90afa","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d8ceb100258db933704ed0ab","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7e2d1d75764c3a08dff9749761f3bc9641c6c1f6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1807,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1807,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-52864107e5dcef4a6415","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e7e30598dd7621c8190a12b5","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":16.317131044387818,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"97d9b75127c8cfd4abb81ec78ff776b43469e1cb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":647,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=39;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":647,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":16.317131044387818,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2c5bdec2880d56a72050","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-fb9c98c8f055c1e29d4a7edb","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1835.337,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c938b04cfcd1a01f543b9d49ea069776e8027404","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1811,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1811,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1835.337,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ce034e417d58e6dccdf8","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ec1e9933b7ec6965f274ec63","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd8ea9d74c7d498769469179a0a6d9b620009493","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1810,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1810,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-db42810116466fb619b1","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-97489ef39730547c2dffe7f8","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d8e9738c5955d63927577a1d1d9e4e8ae7fdc10f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1808,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1808,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9b9794a05018a03b2e04","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-abffd643db0d91bfd1b01d44","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6360fe817a8c393af223780c324276cfab1876a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1809,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=39;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1809,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-660cd060a106965f0ef6","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0da8006ffa1e5ddd8e2fcec0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Haiku 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Haiku 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku (20251001)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.8390833333333435,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2af6c29d942485b6eaee541bad19d6d5d90cacd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":239,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=39;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":239,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8390833333333435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-55bda297f0c5969f78a9","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-9c678bffa01049fc61e41a20","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0d66e0d766f0ae30be9e7ed8eb6ee6cda1f3445","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1782,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1782,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d8651eabe24fe963f5f2","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-11c70a0e094768c442658f62","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":38.54364204096484,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21d80feea4226d95da690eac5529a38ba5a3b3d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":641,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=38;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":641,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":38.54364204096484,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-a98b1a2b856c5c5262b2","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6acf50b7b4ca2eb4119066d0","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1804.604,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87d8a00ba15cd1c5e7e1c4ca0f3e3ff1bf7cb67e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1786,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1786,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1804.604,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-cf6ef330a62ac3d6a1f1","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f02b705c59c7ce376eebcbed","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf98c51ae28b58f2fdee93276bb5ee4b494dbc6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1785,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1785,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-e414fa47f7e99dbd84bf","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-80737d933d3bef5d085a619f","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7cf272483d6e2c52156ab21250a0c74e1c8efdb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1783,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1783,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-7815d7359eaefeacd3d5","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-90c9c451a401216c5da03bbf","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5bfb09883ffe9d33d2609cef727235eed1d6c575","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1784,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=38;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1784,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0ea65a67296f3da9af2a","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-903b914a0bbc990b64eb61b4","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"HELM label for the dated Claude Sonnet 4.5 release.","registry":"model_aliases.json"},"mappingNote":"HELM label for the dated Claude Sonnet 4.5 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet (20250929)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.8544444444444534,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0646076e75a48e54212b97e57a6befcccee32552","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":233,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=38;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":233,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8544444444444534,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ebdb0c0dddb6755eff31","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-81bfea7d37f030899f55db3d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21b53be129bb75ca93db67751c4c0a935ee926b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1457,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1457,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4e3ce755af720f9b7073","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ae1f8918a6e0b131ae0c6a52","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":87.84843708276749,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e4073ae938ef894cc8eaa310d935d93bfda9789","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":563,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=25;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":563,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":87.84843708276749,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-752c06b41e1341d04f5f","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-15613943690143fe02a7039a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f723b78a54cc4211a53f333d7797062fb0a36b5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1461,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1461,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5be62504a0ece5541f2a","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-61a8c7bf1a107cf8ce104f89","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"16c437e782cd64416fcd3af08f5152c2f43feeb6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1460,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1460,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c514305b5106bf0c179e","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b0346c2d6d2fc2369b092ec9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9de80e33b84f063ba83b1fc8ab9441953cb564d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1458,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1458,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c98cd3bd7e50fc678803","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ce41093fad63a1ce1c342312","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"680f9735eeeeab53604064adb34cd10a7e5ce5ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":1459,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=25;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":1459,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2cc06545cab73236521f","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4f25c3d88b94d2767cd2ce1d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":0.8280925925926018,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d37fef5045a4bf5f125d267fd9b3fcca31a5b1a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":155,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=25;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":155,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8280925925926018,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-249f5b09184584418fae","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0a41b8d6fc502331befe7d48","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a62fb4ce5e29ccbe46db707c13e20327773b750","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2207,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2207,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d46b3f40565db47abb87","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-6f98a3cd87d62fa572260f3d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":88.59490567517281,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"121d002de363b8b93f676e74524181de3445d304","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":743,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=55;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":743,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":88.59490567517281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c4b4bca46f9f8d484673","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a049bf4dde61407eb8b85f7a","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1518.974,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c4e2a9908faf936e450883017e3e7228eeaa6ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2211,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2211,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1518.974,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-4c7b4e28e4cb6f7ee9f7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-b9b0ba6715f62c6bfcf18adc","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c6b7530d9dac28437d3811eb8166bc5eb8ac58e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2210,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2210,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-f731123ecb057b6c7ba2","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-5c25cb61e832388c667d0dcc","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4282f519540211efc4feece36a9f75f835214dd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2208,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2208,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d847b2cba8464afd331d","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c3e75a81d0fc6e076f5fac81","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1692053912ff7852d8ed490c960d751cbff6987d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2209,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=55;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2209,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c920ab4cc16c53b13963","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-af0f5da55fd755f4e42931b6","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"GPT-5","modelRef":"GPT-5 (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.8573518518518621,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"63383c59e70c09447ec39551fa42a0739e93ecbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":335,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=55;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":335,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8573518518518621,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0d4f2cc30a17654c1f3b","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-07b78eef4757d87bc213061a","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b5cad07bfb86c29d744a0e6305911ab2207ee31","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2232,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2232,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-0d8fbcc113a84ff4ffe7","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-12fd94cdd5cccafa12683d74","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":33.85394237089157,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b376beca90651d5414e1524a644b7dc4a5e3697a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":749,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=56;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":749,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":33.85394237089157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3b9da2d4693b8fd7fa59","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-a8106697f39f08f1cdb8d427","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1408.024,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74b5d8cb7903c20adc859e576dd3b70cfaf375ae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2236,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2236,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1408.024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9c5de8c568ea52d162d2","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-c5f4a74cb43992315e72fe9d","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a430fd2328db2b73cd02742df7cc95a33bc27120","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2235,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2235,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-023cb8ae28236a1e8a3a","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-39c725b030e2ec3bc3b7d55e","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72bac70e5fa4ac8692a1160f947a6a466d87c822","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2233,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2233,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-c997aea06c50bdade1fc","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-0e44844d778423e1bafad80c","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d83af99e509c28fc1455bb553a403358f2dbe81","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2234,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=56;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2234,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-d63045364c07ea3d57eb","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-ab355b0db1b38766c085e73f","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"GPT-5 mini","modelRef":"GPT-5 mini (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.855296296296306,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1edb50f31ad6922df948a4c274bca105f0040dc2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":341,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=56;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":341,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.855296296296306,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-ba5a3eb0e09a13466168","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-74e1ad2a9a64064397e7c5a8","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11e12fb8a1dc5fe50d619fddb0e95753287e4618","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2257,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2257,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-9a628ac8df9a77510c97","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-2eebfd77e62ea86024f00fc8","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":47.560468022584914,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9630c697bc5ee79bb814dabe0ccc5ba415e92ada","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":755,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=57;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":755,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":47.560468022584914,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1d1842f9d7061621960e","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-23132c3f213d935ceb540988","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":1391.144,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"09d843b86b2e40625799eb5361dc3e6b01dd5518","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2261,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2261,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1391.144,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-dc9c275a34032f3e2235","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-1145ae93415c18a5fe74c232","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b4cbb56b3d54235a09600577958946c83e92c08","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2260,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2260,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-e0519c7f994c7b5516f3","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-8e380fcc6f8e5052fc18a557","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89b1fc86087e7af029cc2f9547952f5c1838cedc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2258,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2258,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-5aed0210727dd63f925e","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-3cdcebf6ea1f2d850fcc2385","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d4d4e23a853b5e97e40894f900020ce3bda2fba3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2259,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=57;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2259,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2c8fc6f443fb5bc7124c","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-958e5cd3e86108a1de7c50ba","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"GPT-5 nano","modelRef":"GPT-5 nano (2025-08-07)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":0.8059074074074162,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f094828b517b73c7137dfb0d1cfca36164280632","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":347,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=57;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":347,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8059074074074162,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-1f9492edc5091e42142a","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-23a7cc01cbd510c178261925","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"eval","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":1000.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d434e561474c8d4f82d65887c57675a30675b6e4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2282,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=WildBench - # eval","sourceMetricId":"eval","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2282,"metadata":{"direction":"lower","header":"WildBench - # eval","header_description":"WildBench\n\n# eval: Number of evaluation instances.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":1000.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-3c52ef93846e72b0d531","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-d3bb9d4352481476b10d25c9","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"observed-inference-time-s","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Efficiency"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":28.08133857488632,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc86e10fe2c349fd4369cec68c6f718235838729","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":761,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Efficiency;row=58;column=WildBench - Observed inference time (s)","sourceMetricId":"observed-inference-time-s","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":761,"metadata":{"direction":"lower","header":"WildBench - Observed inference time (s)","header_description":"WildBench\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"seconds","value":28.08133857488632,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2c81b0570f38bccd4e4e","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4d6b94988ac221e609252f96","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"output-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":2059.716,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bd7b5056d6deadcdbe48297b16ff2c2988c24b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2286,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=WildBench - # output tokens","sourceMetricId":"output-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2286,"metadata":{"direction":"lower","header":"WildBench - # output tokens","header_description":"WildBench\n\n# output tokens: Actual number of output tokens.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"count","value":2059.716,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-afa9bde13733d39f1acd","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-e4e8a7be64bebaab43d62e2f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"prompt-tokens","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b1349df7bbfd785e9b773a0dac3ca1074aa58a2c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2285,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=WildBench - # prompt tokens","sourceMetricId":"prompt-tokens","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2285,"metadata":{"direction":"lower","header":"WildBench - # prompt tokens","header_description":"WildBench\n\n# prompt tokens: Number of tokens in the prompt.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-82267b1b0584c726f0e4","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-f9201e418cf2b076c1431d0f","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"train","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"lower","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"811955adbdae9d78dea6761ab073b83621cf0ce1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2283,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=WildBench - # train","sourceMetricId":"train","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2283,"metadata":{"direction":"lower","header":"WildBench - # train","header_description":"WildBench\n\n# train: Number of training instances (e.g., in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-2a0258a894caa0c22e38","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-4588cb7abe23a30456032d7d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":true,"matrixExcludedReason":"telemetry_metric","metricId":"truncated","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"General information"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.0,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5385c3e9321b4b7b5bbda683ea2c98883ecf951c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":2284,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=General information;row=58;column=WildBench - truncated","sourceMetricId":"truncated","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":2284,"metadata":{"direction":"higher","header":"WildBench - truncated","header_description":"WildBench\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"helm-wildbench","benchmarkName":null,"benchmarkVersion":"v1.15.0","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-helm-capabilities-fae5e87082f7e727acec","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","harness":"helm","harnessId":"helm","httpStatus":200,"id":"pub-953a4e2cbdc540707460fcf1","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"HELM release label for GPT-5.1.","registry":"model_aliases.json"},"mappingNote":"HELM release label for GPT-5.1.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"wb-score","modelName":"GPT-5.1","modelRef":"GPT-5.1 (2025-11-13)","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","protocol":{"direction":"higher","harness":"helm","project":"capabilities","release":"v1.15.0","source_status":"published_release","table":"Accuracy"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":0.8629444444444558,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9380363be92e8aada4990c47634c707ada832247","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"helm-capabilities/candidates.jsonl","line":353,"payloadSha256":"63e1661cf2158549e56da402bec46d0be439959a8db8a78603da28d3543bc04c","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"helm-capabilities","sourceLabel":"Stanford HELM · capabilities release artifact","sourceLocator":"release=v1.15.0;table=Accuracy;row=58;column=WildBench - WB Score","sourceMetricId":"wb-score","sourcePageUrl":"https://crfm.stanford.edu/helm/capabilities/latest/config.js","sourceRow":{"artifact":"helm-capabilities/candidates.jsonl","line":353,"metadata":{"direction":"higher","header":"WildBench - WB Score","header_description":"WildBench\n\nWildBench Score: Score of the AI output judged by GPT-4o, rescaled to be between 0 and 1.","project":"capabilities","release":"v1.15.0"},"sourceStatus":"candidate"},"sourceSubjectType":null,"sourceUrl":"https://storage.googleapis.com/crfm-helm-public/capabilities/benchmark_output/releases/v1.15.0/groups/core_scenarios.json","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"fraction","value":0.8629444444444558,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-89602072cfe66afc17ea","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-cbc0d8e327c2215ef8546c4a","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.1","modelRef":"MiniMaxAI/MiniMax-M2.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":58,"rawValue":22.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a677026a4723fbb8a3420e64a2a0243826a94e49","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":58,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=58;row=57","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":58,"metadata":{"author":"MiniMaxAI","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":22,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":22.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-31c09d5d208cf779bea0","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-86dfdca7e71d74cc14b16f12","mappingCandidates":["minimax/m2.5@2026-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.5","modelRef":"MiniMaxAI/MiniMax-M2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":64,"rawValue":19.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06aa4c5ad5e44ad7c576f594c0a40ef1492684dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":64,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=64;row=63","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":64,"metadata":{"author":"MiniMaxAI","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":228703644928,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":19.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-00b1c1e0435571d66e40","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-0fbcbaa0d06847989ad31990","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools_4.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":17,"rawValue":47.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8d4234b6d409f56642b99200001ea157e8518616","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":17,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=17;row=16","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":17,"metadata":{"author":"Qwen","filename":".eval_results/hle_with_tools_4.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":8,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-c02005e06f0716306c98","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4c0e8f9762c828a6a5c00232","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_chain_of_thought_3.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":45,"rawValue":25.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8d4234b6d409f56642b99200001ea157e8518616","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":45,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=45;row=44","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":45,"metadata":{"author":"Qwen","filename":".eval_results/hle_chain_of_thought_3.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":8,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":25.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-afdfb2699a9dfe431b2b","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-edf4a2019a2785baf73cd3a4","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools_3.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":12,"rawValue":48.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed4b9d984c5210881d4b3ec6b24441682fa45031","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":12,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=12;row=11","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":12,"metadata":{"author":"Qwen","filename":".eval_results/hle_with_tools_3.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-847f05f9ed3f1ac70044","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-57b97116f5fae7ff54a8f7ed","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_chain_of_thought_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":48,"rawValue":24.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed4b9d984c5210881d4b3ec6b24441682fa45031","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":48,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=48;row=47","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":48,"metadata":{"author":"Qwen","filename":".eval_results/hle_chain_of_thought_2.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":24.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-c119334ccedad070472d","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a078e0829bd60ad12ee0d6dc","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":48.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d707a2678da2787bd39295ddb066a0a190db2455","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":13,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":13,"metadata":{"author":"Qwen","filename":".eval_results/hle_with_tools.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":9,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-67a2951b48d97b07c218","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-aec354ba47bfab112a7b1f11","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":40,"rawValue":28.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d707a2678da2787bd39295ddb066a0a190db2455","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":40,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=40;row=39","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":40,"metadata":{"author":"Qwen","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":9,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":28.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-9cb0a9c35880cc3d6e99","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-62e0f19528bf40bf93428b2a","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":49,"rawValue":24,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a860ac872ff51b79493f88796dc03e581dae6be3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":49,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=49;row=48","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":49,"metadata":{"author":"Qwen","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":24.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-97b4a91c09b195e6e8de","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-3938dfac52cf6baa0eef0f83","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":62,"rawValue":21.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e2af8bcc1e902a90db25c00205064c1fa6106a58","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":62,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=62;row=61","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":62,"metadata":{"author":"Qwen","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":21.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-89f02d1e28b22cf3f9ef","canonicalModelId":"qwen/qwen3.8-2.4t-a95b@2026-08-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-e0521489452a41891053df36","mappingCandidates":["qwen/qwen3.8-2.4t-a95b@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 2.4T A95B","modelRef":"Qwen/Qwen3.8-2.4T-A95B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":21,"rawValue":43.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25d48fbe576ae2042c9254877541be916f4dc983","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":21,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=21;row=20","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":21,"metadata":{"author":"Qwen","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":2446182725504,"pull_request":6,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-21edce6867b9bfeee13f","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f1079df2a35bdd2c40e1655b","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"Qwen/Qwen3.8-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/Qwen3.8-27B.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":35,"rawValue":30.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e487378338df400a0120eb79ebdc14672ce6bf0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":35,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=35;row=34","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":35,"metadata":{"author":"Qwen","filename":".eval_results/Qwen3.8-27B.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":22,"source_name":"Qwen3.8-27B model card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":30.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-24b4b1db036abb7ee34c","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-88d9ed2df90763666f3c0c43","mappingCandidates":["qwen/qwen3.8-flash-next@2026-08-26"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"Qwen/Qwen3.8-Flash-Next","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/Qwen3.8-Flash-Next.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":27,"rawValue":35.9,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e72f5ee31d918b02865310a2d0cbd9f06a565f42","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":27,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=27;row=26","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":27,"metadata":{"author":"Qwen","filename":".eval_results/Qwen3.8-Flash-Next.yaml","lower_is_better":false,"num_parameters":179999981459,"pull_request":12,"source_name":"Qwen3.8-Flash-Next model card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":35.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-86ba51e71faa140c19b0","canonicalModelId":"xiaomi/mimo-v2-flash@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d8782ae0c844eddcf649195a","mappingCandidates":["xiaomi/mimo-v2-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo V2 Flash","modelRef":"XiaomiMiMo/MiMo-V2-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":59,"rawValue":22.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"791e19da57d885092d5aeb7799799ee643bbf296","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":59,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=59;row=58","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":59,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":309785318400,"pull_request":31,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":22.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-002024492da3e36ea256","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-08bf7297e416b9cb93a01a21","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5-Pro","modelRef":"XiaomiMiMo/MiMo-V2.5-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":15,"rawValue":48,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ddf2ae812a2fc8bbc6053a38c8e1ec5e9abe4420","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":15,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=15;row=14","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":15,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":1023244718976,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cf3e1a00766d7399b0e7","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-cc45ce0825586adb23027cee","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0668","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e5be137d8fb111dd97c88dbd64c5a76b5a4124e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3301,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=41;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3301,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.68,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9b6802404f8044273a21","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b698653ee5225f22d38ea160","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514_unknown","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.1072","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e5be137d8fb111dd97c88dbd64c5a76b5a4124e","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3290,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=30;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3290,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":10.72,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-58340fe8e30071639ef8","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3a41e0e83caf8f078955759d","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.1416","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376d773a3f15c4e54b7f46190bca45a89267b042","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3285,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=25;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3285,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":14.16,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-70739c2c5a9149fb0f0d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-36b7aa176d1a808b1875ea1a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_unknown","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.252","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"376d773a3f15c4e54b7f46190bca45a89267b042","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3273,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=13;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3273,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e6b0732f219d98323607","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-eda40c62418b192d9a35a262","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.19","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5354ec76660c79f52ca42562e0b70131340779ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3280,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=20;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3280,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0fb8d5df55ad5cac3247","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-623af3eab03cc59f08d007c5","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3444","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5354ec76660c79f52ca42562e0b70131340779ff","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3268,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=8;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3268,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4785cbc2829b0a8ec334","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ddd92195cceff27a61f5321c","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only on Claude Opus 4.7.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only on Claude Opus 4.7.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_unknown","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.362","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82bc2c6420b318f7a0732777bb2c59d03f1f3440","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3267,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=7;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3267,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.199999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7ba577e8715644e1ee53","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4f216dfade4ce70f7bf74bf7","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.0752","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a48590ce481fbee21209119fe1391af40a8e9f1a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3299,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=39;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3299,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":7.5200000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e375ee942e571285c247","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1ab8a110fcd273f0a73db7d2","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_unknown","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.13720000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a48590ce481fbee21209119fe1391af40a8e9f1a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3286,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=26;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3286,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.720000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-b62ef93a0366e80c7bf1","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-58c2f9c488693fc8b274ddf7","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-ai/DeepSeek-V4-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":28,"rawValue":34.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0245e224e9ea556913d6653b3a0dc8e5515a0c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":28,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=28;row=27","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":28,"metadata":{"author":"deepseek-ai","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":290944616402,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":34.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-b4abb67333911fc254eb","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-679f7fd725c79a4d61795f48","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-ai/DeepSeek-V4-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":24,"rawValue":37.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"194622f3dc3fb35c69daf6403ab1e8baf498945f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":24,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=24;row=23","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":24,"metadata":{"author":"deepseek-ai","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":1598839674782,"pull_request":110,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":37.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0b1799edcdda7f5af967","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-dc438fc6e54129fd392825d0","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4644","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b04a5aba9bc82359417940b030fe91208fcd4bb8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3262,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=2;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3262,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7a09de20c09126d463fe","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a6f1b489d59edc2fb3e5cca","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2532","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"163691ab5d4ddd7fec42f1fa2634ced324b66567","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3271,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=11;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3271,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.319999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-51b1fb5248ab2f5bdd91","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e396bf7aaff94421b973a017","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","registry":"model_aliases.json"},"mappingNote":"Epoch setting suffix only; the dated GPT-5 release is explicit in the alias.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2532","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"163691ab5d4ddd7fec42f1fa2634ced324b66567","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3272,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=12;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3272,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.319999999999997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b4292865a7b09e957bb8","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-17bbe28a2c5a1e13a89ebff5","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_unknown","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.19440000000000002","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab2dedf25421b242bc13fe46c9021bf63525b715","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3278,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=18;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3278,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9ef76b655be41b661a86","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aea6ffa429124373d2868e09","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_none","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.068","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d9f15824326f9805a93974b3f378898db060232","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3300,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=40;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3300,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":6.800000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dd4a669d42f59a3e6013","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ffaa5d14b500cc5592dac91c","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_unknown","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2368","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d9f15824326f9805a93974b3f378898db060232","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3275,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=15;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3275,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.68,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-07fc152ea6b679ba652d","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c9f36038c51123d3fcaa13d","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_unknown","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.278","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e6a8508e522eff0ee185512560b6c3e7171d7bb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3270,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=10;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3270,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.800000000000004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-68ec82aa26b1e4d2716a","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-879d4f51aa237bbf985b3f1f","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3624","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"41e48185fdd9de637cd89f67bb67d48d5de426c2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3266,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=6;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3266,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":36.24,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-37a06547f13af04f7ae4","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-92080b202e442249a41f2d68","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_unknown","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4432","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a64e9f2db47983aec8d0d2b3688d583096868b26","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3263,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=3;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3263,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.32,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-aebbc279a97d4afdb873","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fde97f71b881ad4688a9a0a1","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Accuracy","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.24370000000000003","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e832f4a594c950c3baaaddd469b8a948fcb35519","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3274,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"hle_external.csv:row=14;column=Accuracy","sourceMetricId":"Accuracy","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3274,"metadata":{"epoch_file":"hle_external.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":24.370000000000005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-ee7de127c347f3ee4fff","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-9191907b3e8262ab7dc6f92e","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":11,"rawValue":50.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b286969fd2b13718ffd45d3387e694264263e5ee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":11,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=11;row=10","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":11,"metadata":{"author":"moonshotai","filename":".eval_results/hle_with_tools.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":49,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-1d0c57e9f858186d8120","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/aa/05/aa055fe5-5174-4517-8b10-19d46da94f3e.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-c7578c808012322d95ee7aff","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":47,"rawValue":24.37,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b286969fd2b13718ffd45d3387e694264263e5ee","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":47,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=47;row=46","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":47,"metadata":{"author":"moonshotai","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":126,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/aa/05/aa055fe5-5174-4517-8b10-19d46da94f3e.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":24.37,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-f100161edebaaa63a054","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d2793a98cb3d2c41f6c62f46","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":5,"rawValue":54,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7297e0eca0153adc98cf2ed1bbbec87e14238452","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":5,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=5;row=4","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":5,"metadata":{"author":"moonshotai","filename":".eval_results/hle_with_tools.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-a22a7c59b9adad4c4819","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-0f6006e3ecd91151e70d8e5c","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":29,"rawValue":34.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7297e0eca0153adc98cf2ed1bbbec87e14238452","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":29,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=29;row=28","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":29,"metadata":{"author":"moonshotai","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":34.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-85b31636d97980aea157","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K3","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-98af02799329368f475b73df","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"moonshotai/Kimi-K3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":2,"rawValue":56,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3338b48813e9920a9c5e839a5b78ae9a2d2bb5a1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":2,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=2;row=1","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":2,"metadata":{"author":"moonshotai","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":2779931837184,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K3"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-fcc04b9e249237cf35b9","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a9612a5c3c5c046442a41aca","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":22,"rawValue":42.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c5e0f0a2f1f9eb1acf9548cba4e2f1c90e214df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":22,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=22;row=21","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":22,"metadata":{"author":"zai-org","filename":".eval_results/hle_with_tools.yaml","lower_is_better":false,"num_parameters":358337791296,"pull_request":40,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":42.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-2430ef7257babcb5380f","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-177ba21f61bdec464631184e","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":46,"rawValue":24.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c5e0f0a2f1f9eb1acf9548cba4e2f1c90e214df","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":46,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=46;row=45","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":46,"metadata":{"author":"zai-org","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":358337791296,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":24.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-901c8bad470f77548f09","canonicalModelId":"zai/glm-4.7-flash@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-0d0d4e783b767dd6da2389b1","mappingCandidates":["zai/glm-4.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7 Flash","modelRef":"zai-org/GLM-4.7-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":73,"rawValue":14.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7040f93268c258c167753232b8a82a1495ed2f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":73,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=73;row=72","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":73,"metadata":{"author":"zai-org","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":31221488576,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":14.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-45307579476c0209cb28","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-b484de33869a84985d88e436","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":10,"rawValue":50.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2d1462e4a2678e9c26aaf41d1eda8c6824322de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":10,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=10;row=9","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":10,"metadata":{"author":"zai-org","filename":".eval_results/hle_with_tools.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-f82eebc586fb50517678","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-39c6d296a205a0a057b7c5e3","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":36,"rawValue":30.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2d1462e4a2678e9c26aaf41d1eda8c6824322de","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":36,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=36;row=35","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":36,"metadata":{"author":"zai-org","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":30.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-6b003c460f05001b81d4","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-ac845176305cd06ce4019def","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle_with_tools.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":9,"rawValue":52.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"324f9d834943e9a433c06ac29248e60d42f4b141","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":9,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=9;row=8","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":9,"metadata":{"author":"zai-org","filename":".eval_results/hle_with_tools.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-9bd9852a3b2118e4a386","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f08479ac5efda2747ccb4a42","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":34,"rawValue":31,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"324f9d834943e9a433c06ac29248e60d42f4b141","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":34,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=34;row=33","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":34,"metadata":{"author":"zai-org","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":31.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-2c175108f4b2b35cffef","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.2","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d9d97dcd4af4f1d6e37ed767","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"zai-org/GLM-5.2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":4,"rawValue":54.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cfcbb0c82cfad5bd3d70ab50ede0d4b5efc093b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":4,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=4;row=3","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":4,"metadata":{"author":"zai-org","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":753329940480,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.2"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-9b85cb385bbb6c62d51d","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.2","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4476e035caf9caea483a1d94","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"zai-org/GLM-5.2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/hle.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":23,"rawValue":40.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cfcbb0c82cfad5bd3d70ab50ede0d4b5efc093b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":23,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=23;row=22","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":23,"metadata":{"author":"zai-org","filename":".eval_results/hle.yaml","lower_is_better":false,"num_parameters":753329940480,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.2"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":40.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hle","benchmarkName":"Humanity's Last Exam","benchmarkVersion":null,"benchmarkVersionHint":"hle@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hle-2fc0ea93d06e4163cc04","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.3-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-bda8d7fd370b92bdb71eae4c","mappingCandidates":["zai/glm-5.3-flash@2026-08-27"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"zai-org/GLM-5.3-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","protocol":{"filename":".eval_results/GLM-5.3-Flash.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":3,"rawValue":55.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06c0664ed82bd2792fae7cd56716bf6b4be8486f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hle/candidates.jsonl","line":3,"payloadSha256":"2a0ab0a3296a218e1531287ca7eeab13d7863d305a00987333f0880bfa506ae9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hle","sourceLabel":"Hugging Face · Humanity's Last Exam leaderboard API","sourceLocator":"rank=3;row=2","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","sourceRow":{"artifact":"hf-hle/candidates.jsonl","line":3,"metadata":{"author":"zai-org","filename":".eval_results/GLM-5.3-Flash.yaml","lower_is_better":false,"num_parameters":321323031390,"pull_request":12,"source_name":"GLM-5.3-Flash model card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.3-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/cais/hle/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-ffeab4ae75486afbe696","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-6bc206685f5c7bf531b80703","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/MathArena--hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":14,"rawValue":81.06,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9be1dc280f47e60317b5851d4d6b48a70665313b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":14,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=14;row=13","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":14,"metadata":{"author":"Qwen","filename":".eval_results/MathArena--hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":35,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.06,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-971b5c0a45ed0271958e","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-adc4d0c7b4ce08c78e6e1ba8","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/MathArena--hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":4,"rawValue":87.88,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d06ceac6eac1f9e6dffbc07a14227dcd2a77d4c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":4,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=4;row=3","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":4,"metadata":{"author":"Qwen","filename":".eval_results/MathArena--hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":53,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.88,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-cba141b7e19774f005a2","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-35ff96473af6acd55251dfca","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":11,"rawValue":84.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dfd2796536b2c3599b5b50fa36c26fa9dc493d8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":11,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=11;row=10","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":11,"metadata":{"author":"Qwen","filename":".eval_results/hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-cf5ac4d5309714cca98a","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-879e3f654105211d00ac0c1f","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":12,"rawValue":83.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2bc06947ab17961fb293a8cdbf54bb9fe609859a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":12,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=12;row=11","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":12,"metadata":{"author":"Qwen","filename":".eval_results/hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":83.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-d71fbe9b7c5c8adb16f9","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-3c585259627eb607b7f52d8d","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/MathArena--hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":6,"rawValue":87.12,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d759e3fc7597a8bd76d2cafcc7730ada4ec1b38","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":6,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=6;row=5","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":6,"metadata":{"author":"moonshotai","filename":".eval_results/MathArena--hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":95,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-5f90b5b6502551e7352d","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4130b33c5b0429795d94d747","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":1,"rawValue":92.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7882b8111400d8b4577db76278b7a5c59379b893","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":1,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=1;row=0","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":1,"metadata":{"author":"moonshotai","filename":".eval_results/hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-eb18a4cd31ec69301358","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-6d795dd58b1a8914e9497d07","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/MathArena--hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":9,"rawValue":86.36,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b0d97e3717a0129f35a5b7701756823e5c48154","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":9,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=9;row=8","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":9,"metadata":{"author":"zai-org","filename":".eval_results/MathArena--hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Official MathArena Evaluation","source_status":"reported","submission_source_url":"https://matharena.ai/?comp=hmmt--hmmt_feb_2026"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.36,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"hmmt","benchmarkName":"HMMT","benchmarkVersion":null,"benchmarkVersionHint":"hmmt@2026","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-hmmt-2026-4fafbd12eb3efafae750","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-c5c7f45fac43f622863cca57","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","protocol":{"filename":".eval_results/MathArena--hmmt_feb_2026.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":82.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ec350079da3ee94f13ff0c1d23904249c01a960","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-hmmt-2026/candidates.jsonl","line":13,"payloadSha256":"2e159d9614de10ac2db4a6f240abde1aa936df4b4a22fe613d61729786f322b9","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-hmmt-2026","sourceLabel":"Hugging Face · HMMT leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","sourceRow":{"artifact":"hf-hmmt-2026/candidates.jsonl","line":13,"metadata":{"author":"zai-org","filename":".eval_results/MathArena--hmmt_feb_2026.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/MathArena/hmmt_feb_2026/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench","benchmarkName":"LiveBench","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"livebench@rolling","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-56b90ccb80ea2e0e7240","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://livebench.ai/#/","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-671f771e1a662a547f4ffbec","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Global average","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"78.79","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b921ccf2619777529b62a1ba987428a6f975a037","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3468,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"live_bench_external.csv:row=65;column=Global average","sourceMetricId":"Global average","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3468,"metadata":{"epoch_file":"live_bench_external.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":"https://livebench.ai/#/"},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.79,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e6d5bfc806b5646937a6","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fe0c7b0f25a682dda0a61d8f","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25eaaca8c3758fa2b4b4b8ff404b3ffc66b5da83","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":645,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":645,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2021eef60d242c850ea6","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d30cfee397ac8dcfb3cb6b82","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7adcfd910d1a47c9ef646d2f2b772e8d15b94ef4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a83d894fce9d02c47081","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d7ad543ea4026e95955f9396","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d964cef7013c9322cec40768379d4e2aeb8f5827","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":24,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":24,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e167c2bb5006c2184420","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0f2bb829d3ef5531a8267d28","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbc95bc88f3e708f47bd6df41ac41b588f1f90fc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":47,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":47,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-18e059b797cb9753dfe2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0fd86807638348584211091a","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37903f230f541d93b48fc77e95c5de4535fe06da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":806,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":806,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-77a608c76da506071c27","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-84448e42a8ce27866e80f19f","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.01","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bea9f35eddebc57c26bd8e650b3db04b469dc69b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":783,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":783,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.01,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-06aa7e7980996f4ac009","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8186553f7fc718059efd5a76","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bac6d05dfa8659e439346df077d788c928270e4c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":70,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":70,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-82fefe272b7674c31983","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-19dc94f6fae6d881f38d82de","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4744be60f09f3303e9bbc6e56232298d8e56b42a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":93,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":93,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f6e514cdbad433185f85","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f5d56b886a54dc5ab3580d82","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2673965cade5c3503fe22c5888da6edb9d923b45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":116,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":116,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-52cd896cc36ca7df78c0","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b57cf9ddcf4da0c4fd376c04","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2673965cade5c3503fe22c5888da6edb9d923b45","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":829,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":829,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a6b1d9f51ea963330149","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-21983166716d3994cfc832a0","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c164f5becf18b60a9657cca7acc2738970cd3a55","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1036,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1036,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a5caf9d968adf7ab8aa4","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ec688a02af9e8cb56be90a6a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d960ad3b4197c8bb5888cacca9faeabf30ab5a63","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":139,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":139,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7c79dd5974277025445e","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f46c3c8b8d336816aa8d474f","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b6832c6e868e809df05d260d56530a28583ff79","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":921,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":921,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0eac9b06d1ebe6c85be8","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-50b09b69a69fdb380e83df08","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35cbee0cf50047850f7933cdef7113869c211f43","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":162,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":162,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8adbdaef442e935ed25a","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fe4c40feb79ddf8406e6e6d1","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e822813ece78e26babbec1ddaf117dff4d0681a7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":185,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":185,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f7905ee817bfccc2a9c4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d6010ff95d0270bbec6f9573","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1c45e36b47d5504763b299ad698c9ce28801dd8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":737,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":737,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c94b2b20391708c6af2c","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5e84c5ee11e477666e0d6f52","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4464bbc1180342f7c6538073fe4cce992cd82a11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":760,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":760,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-52d2bf957876d99b6274","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a5c753fa8d725b0eaba6dd56","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f8d200db36f90e1e5c7ddca2cc8a93d3f7d6e853","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":967,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":967,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-44efee9974e35afe8079","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9864b0e8a42558a99ef73c8a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06a80410b09a30de0eda77cff1379b64208288df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":208,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":208,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b22fb962b87f99b46426","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f50d17661bea3e4a2c02bab0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74fce3138de2cf8f1f3b3d0cc17baf8d4426d3db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1059,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1059,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c070046a61cb4fc7d0bd","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fbd66e4d79585c998a4051aa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e994d3560561c15bbd52c9ae438d0b42369fc0ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1082,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1082,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0b93834cd47167014b03","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-02ddfeaf743ce4b75817a8d1","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c755f0b7f864a73df68fb24e940e3334d956c4e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":231,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":231,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e2e4a8cd7b480358fc8a","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-54836708455af060a2eca601","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a53888c371c9fd7129853fffc31c2a0cc0022b27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":254,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":254,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3fb27c8f02cab573e12b","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-16e11253a7b4d134b3bdbb53","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"864f14c9152493e6b8f8349e023dc99fc5c9a6c9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":277,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":277,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-78c85e3a6c712373a959","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-17749dd2aba21d80d8643ccb","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58358fe34730ff5c61f27641e641aa4367041e8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":300,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":300,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-80ab01c7c8e245a30856","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-25a37595cbfa6db5967b54b0","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3bad2ac68edff03ee44feda2eefc7c7db41d58d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":323,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":323,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6d94d51c0b5ea9416fa1","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-63ba99084651cc0fa8b430d5","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fc5fe0c5f4d885af41b40729eb5db5800429030","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":346,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":346,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-da32069b63d2da5a7cec","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9fe3fe17b82dd04e7e475404","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fdedffef634c8c1b29b073a127d803ca7553fa01","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":622,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":622,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-66bc306794112a3b8140","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-59443cd2e4c848b2908c8170","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5c60122393d457be3fa450750ffdcdf197508fc9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":576,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":576,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8a49c3a84e30491c360d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-71072897943593d9469ca617","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef467eec7d14df5cbd241bfddce92f5979e1dc7e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":599,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":599,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f60cc18cb9d89cabde5d","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0a60b164fa1ab551270245b3","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0543a3d21c2bbe7dd7b1553a8f83709165bb77fd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":392,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":392,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd66c592d744c3418717","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-57c7de734667216931b801d9","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f21c0575712ac1d9ce527c3a30aaba21e67506cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":944,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":944,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f7498567e81d5cded9e8","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ed19b98c8f82fa3fe5ab7265","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55428f2d855c996609e9b95a559c641a36ef197d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":438,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":438,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1604fc6aa2071611d5d6","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-aafb638b8aa95e485c8432db","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c26842bee12ee16e969b792c50a12a23a35ada25","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":461,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":461,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d0a9e6e3ed1f8ea0afb8","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4e9164d884eae426c9e0046b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"742d9458c2338ddfbdf347a36696d49c1ce886dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":691,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":691,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e31fe59d9a7c92258cc0","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-53c6e1eccbfad3ccc3e9f39b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3421cf50d6b5d186c50c35785526a7bfc717ab25","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":484,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":484,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-97362096020a108ee748","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d203a41f91858d704080043f","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"93.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"48639a49399cdb8c23176dc3308c3602902d1f0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":507,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":507,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6c2f22e854328c29e395","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-03d040180cc32144d034d225","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"408a73d26c63a53824ffbd780c0622f4b9e3d1a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":530,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":530,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7af73c9e5e9908910310","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b83652be7babd625afe50186","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d866c0768d287ffb21bc1a278dc92b8eb15a7266","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":553,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":553,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-be86628e634ea6b27670","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b6614c8a66c0e383ca00e918","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5831b6dd91d3495c2339c2354f027c5f431ac4de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":990,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":990,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6e16340730043b5e29cf","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-32cdeeadc1963e7ec0a33bd7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fefd2223fe6bf111457cf5ce7957a2c7daf4f20d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1105,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1105,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-amps-hard","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6e9903a6b1614209428c","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8dcda60d8f0f5c9aceec395a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"AMPS_Hard"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"765541b0060cf2fe21b9143b697eb4b611d65307","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":852,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=AMPS_Hard","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":852,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"AMPS_Hard"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1c2344923aa0f3e431ee","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-db4f07708e838a2fb9d6f680","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c1a3450292647b5aa63605cc3a14259bae7bdb86","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":646,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":646,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8f9019301aba8a64d656","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c566a313b03e38087d7d9812","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34a23ce6e0435ea6f96bb6efb88e4fd544e2d57c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":2,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":2,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7261d24e7cb8cbeebc45","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1e1fb13c437cfe17bd5535ce","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a6e4e8c4d482d1280d2a4245a01cc9ac4f6adbd0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":25,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":25,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3138d381eafc04a1773e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-525387b122a6d83656ac6cfd","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0be3912ce64d34721b13296a2784346c61a72e01","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":48,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":48,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9a45e5090d09209dfd17","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-13f33163c6aa6650388b5272","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.783","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fbc7f721cfa8af38283388d62c733f17984c7376","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":807,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":807,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d14945cc6e4ee5346bab","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-670e9a180fe85f1a8f843c69","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"82.609","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58707c90b45a8f176567e2ce3b1f9ab94259c4ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":784,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":784,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.609,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-591986434c5b6c0b89b1","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-19ea559be38d14af1a9fcc34","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d737544895f32d7381fb84c5ff9f7af155fc4cf9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":71,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":71,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dd95a00b3e67929d3c7e","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-57667621e01da443858f4be9","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb87d6178cc3b2290686e9217b608cb1bdc068e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":94,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":94,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-10d95a192b181291fbc0","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-55d6726206bf455542d98bed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"65.217","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6e8c82e2165e4dc97af32cca40482b26d9225f22","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":117,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":117,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.217,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9ebdfd18a714ddcf6bd2","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9e43894416c2ad2f3de362ef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.913","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6e8c82e2165e4dc97af32cca40482b26d9225f22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":830,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":830,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.913,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e52b4d31d3de6b5f87a2","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1b2955186b8503bbb09ee6e0","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.391","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce25a03e517ebb2818b036ce52ef866cd1c2889c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1037,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1037,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.391,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-12b4518b88259531ec3c","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5361e54757a9a17fb0f6aa00","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.565","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cfe7010a74691fac63a547ae5ded5d458ed7ca8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":140,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":140,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2d298716e4d9db403d2a","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d7c699041b3a66caca16922c","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e2dd0b24ba7fbfb6ea668139318df1b150013ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":922,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":922,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-24db5dbed211eee432ae","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1c023c5b3848f434fa85042c","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d73dacced9efa6fccc532b62b874ee125d8c2764","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":163,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":163,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c196e791706ab247ec86","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9bc1e88af7424c601f0ed5db","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"80981a68a14c47877c6ec5300b3f3969f7b28656","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":186,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":186,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3f14c52aec35a72ef083","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ade1a17605018e07ea6cc549","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00de8812113b125c6bf825010d1729b4a7f3bf23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":738,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":738,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-be6490de7abce376a8b1","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b73292427bff99a1ae6a0dea","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"578ad02f769b52c580d8cce1435481ba517715bf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":761,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":761,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6af187ffc9b0e5f0c5c0","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e50206d75bb12509aa74d4ac","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3863c889af8f0ff958e92eb7c50274d28b745dbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":968,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":968,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4902626b6175d5e476bd","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e868a3a5d65bae8326d8ef47","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7aed9b13c815ae87801ba0f4180f10b6c441ea2a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":209,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":209,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1c5733e3b2543b85d827","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7a4302e4def7e7133ff1914a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b71eb3cc0bd3b706ba9301f8d6578ec75c37fb9c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1060,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1060,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-274ebecb2eb7eddc463c","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7e9f0289888d44a604991dca","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cbcca46dfa03c6be799ae7a1c01f8fbc6512ba6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1083,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1083,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c837938089725f7c9745","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-74e807fa7a7a727e10b85bce","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"525d797d17524148d075b60af43e3d5a40ad0fea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":232,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":232,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-95cc57a1c2cbca44e51f","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1e12478228fd1b2101b6959b","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"86.957","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4166a4fa66cecfd0938599ae192810f3e383f045","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":255,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":255,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.957,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-936ee3a25008838041ed","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-97a4e64fcb592719476d09e9","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"75.556","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cdcccd6ae7697c0bed84dd4660299f2c3fff5c91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":278,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":278,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.556,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-84f7297d9b7aba69af97","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7721d29544e8285b3ed4e541","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.391","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4444e8ea865bf6d2912d34ca5e1d128891033f73","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":301,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":301,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.391,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-80d6c2cbe0a22f24370c","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a87e1fd4c84b824b65e2973f","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"38505abe02f6f7d5aac764f87b9d103e6a389d7a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":324,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":324,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cb63d91245ecf7b16818","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-693be8360cd801e5c5849159","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"82.609","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d52511d5340924a413f0d1008d802861959e94c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":347,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":347,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.609,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e646e8a36b4e69c13509","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2809a0d0a6e5b0b6e07c1347","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.957","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"688070508717fd8cbfa493c73eb6c096ea4bbf15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":623,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":623,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.957,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-38a888dfd0b230425b0e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-407657160f70f091db686ee4","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.783","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a46351e201ec418f5fe4f1b9ed99e113d160d50c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":577,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":577,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-741e5ca35d754aff5fd2","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-818d4679b5b6f54d55a4ab7e","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.435","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b4b7cb67a2ac3cae50443881bf0d81c9043b9b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":600,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":600,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.435,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4145120ebd09ed81f760","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4d6def8e9564d0c0b2f92866","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"69.565","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8343495d8b6cfd2542713a9908408df3c8e477f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":393,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":393,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f012dd9b4e7aebaced7e","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6694c5c628b8f8fca6d071a6","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"833441ffb0dd9b7dba337836c35f01f7419f686b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":945,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":945,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fe79310ec9061c71aaa9","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-63ef6a1f54d10805aceae688","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.261","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83bab8ec77664e85d7d379d6dd9fc81606efe3d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":439,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":439,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.261,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8231c8638ba6162081bb","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ae4677b451332091006b5316","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f610cd58ab4db0d6a19e0039160e87b0f804a331","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":462,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":462,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3ad6518017b9460e0f92","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2dc64b7fb55db29cb3c82e4d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"82.609","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e597d44d8fbed2f8fb6d08f250216315e70f6fc4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":692,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":692,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.609,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-36764310193e038caff4","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a6983a4a8816ef8b00bfc20f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"67.391","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd2567350da65143e59919948408711d92124ff5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":485,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":485,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.391,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-816b13396ba4151b5b04","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ff166da40ce5113ee7b276ac","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"71.739","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78df46ddab85bcc37020be0fac01aa41a48583dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":508,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":508,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.739,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-677062cc308855fa3ca2","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ac6e244613438b8cb9186faf","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94ac2b95256b43ef919676629a0d7739d6cc9714","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":531,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":531,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5e07ac40dacb1fb1829d","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f70c68df0b2b84d46902e955","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"69.565","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0ba9ed58db06dd5d4a000b6973b671c20246486","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":554,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":554,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.565,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-660cf2acb2ebfe0ee97c","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2c8fc988f4b1a4e7451140ea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.913","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2174efeace1946a0114ac73191bba9379563d798","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":991,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":991,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.913,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ccd5800681e7f953d9fc","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-43bf7f2c5b9592b31460fbc8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.087","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0783aec807ac86988cc663bb26e6f5dd034c3f91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1106,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1106,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.087,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-completion","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-50fadfde74c61cd162f1","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bced40a0ab472e25c0b89801","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_completion"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.913","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"020febeaafae9ab25f84dc56d4c4a07b4e3a5908","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":853,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=code_completion","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":853,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_completion"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.913,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c651fe8f4a18e48ca600","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e1427f074185de5df30ae05a","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.549","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"75a350f85e8d93c506a1392a2af59c987b4a85da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":647,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":647,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.549,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-461bce99df38aafbab7c","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ebf81def3df1e7816ef4a345","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.873","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5d72aab38035c88efcdd81126fb340f0909a802e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":3,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":3,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4b39b60ec783883cd44e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b58b673514e680eb4bef3cfc","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7f951462a9fdc7535eed55707f8c8d080c2ac91","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":26,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":26,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e939a87169878ae44c04","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0d235c18737ed5f3df53b2a5","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"85.915","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce29a41b9cfcc7d80e09ac1e438fcceade5e9fa5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":49,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":49,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.915,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-12f6c54d5e9a0f32ad14","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2c33ba50c5f89dea55133af8","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.873","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29a585857eeba60a8069b2047056741dc5bb68bf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":808,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":808,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-607ff60b993f4ca9079d","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6bd5e4feb327faa93afb9bf1","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e40c86ddedb3dd5f54138272026ec3a6bd0545a4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":785,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":785,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b276def9ca258dc912f1","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-aba0531146fbd6deeed7b288","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74740fb521ec560edacf2b6cdb200facf4f74b8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":72,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":72,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9c2685cb3aa97cb7d613","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3b98cf76dd8749a681461482","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"83.099","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"97d9221ce30d9f46b4c5bd2643d6767d5949e037","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":95,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":95,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":83.099,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3c99ae0c4cfa1a7db6b1","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f118f6207fa3fb44a2c58d0f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.239","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25a0fd31185e6b7b70351b019d8295f721460231","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":118,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":118,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-42e9e625c3daafeeac3f","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-862b84db833c0da21febeec0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.056","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25a0fd31185e6b7b70351b019d8295f721460231","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":831,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":831,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.056,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8a6e0c8ceb4536ca1f21","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b10cd2c58400688d9a6cfeb3","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"69.014","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f649eec4f47e6ca5050441591b437661a4a31878","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1038,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1038,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.014,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e6eb89927cc366149165","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a94284400271436a5a772c9a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.423","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b1e7c0efa6d73dd9f63a0d2e8b26fc033812695","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":141,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":141,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.423,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-21130daa370205266d17","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-05a45ea9110d57a46420f026","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.056","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c78fb16f39137e16a0df7d5b37fcaf28316a3ab5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":923,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":923,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.056,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-762d217452290a482eab","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2dfb7467f2889008f645ab6b","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"74.648","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f789c78677dd8f9468302cb78176c0de567262d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":164,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":164,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3558453571d9a920ed38","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-254f8cf1da07739d0553efbf","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28cb5a4fb76c6578a36a32b22c351836614f3bef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":187,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":187,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2594e36550dfca59d41a","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-92829d3f7ff9db77536832ca","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.056","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"15f1c191fb03031e8113c7b2c3edf7ade1b7f44c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":739,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":739,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.056,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5648f0d0ebb25684bb7f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e71b3ef5996cf04b52d0e502","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"77.465","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"48d6566282664107b55e147fe1bbea17ce6cfb90","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":762,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":762,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-10c0666647e51bed2212","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e6f1c7a712c94d21aa4055c1","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"81.69","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fbfbc594071d16230bf7299400749ae5ea0281cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":969,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":969,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.69,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6ad6b2cba3c065973c08","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d1cd55da8476bc7a821980a4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"78.873","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"569dadd519861a2e0a9651de8203b7ae5e3c93a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":210,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":210,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-94813675d08d500ecbc4","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4e28f68b1391d10329bbb3fa","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"77.465","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d0829811517af1f2d9b443550a2c2a21c274ba6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1061,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1061,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3507f5751fce81c582c4","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d7002b941cf72ceadc4b21c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"77.465","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b423ef7ee4591069f9cb9645504f489ac5c66f15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1084,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1084,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-99ed3c6ae39c8ffc3cf4","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ab0bdcad2232e7ae21a43ef6","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.056","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"368dded365d4c95f5f05f5307d8746be17ee4a64","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":233,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":233,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.056,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ba0044648b0320da4a36","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dd65cc71862f092c41376257","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40f9fef98f08006c1ef9384d7ef26da7dd04b9da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":256,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":256,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c5e709f2c6c2c3111457","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-257282bed877178ed5d11d4d","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"67.692","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6750ce73121db4e0017736cce9713cf9157e6906","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":279,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":279,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-299f96bf788781a8b737","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0c17b6fab10133d6d0849779","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.286","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"210c9b51669806135c36cba96c619cd85f291422","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":302,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":302,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a5e9bd9b173c2895bef0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dfa0b9108366e6f5b5d144ee","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"74.648","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c03b0c448a17ab1d6e2804697afe14b7608a36fb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":325,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":325,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.648,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-acc8df9dcd38b78f9d25","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ef7628b4fb5c80e903e1ca6e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"81.69","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5c6bcc75758031aab3fb6c5ec1fd7b9bb1a7121","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":348,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":348,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.69,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6c6bf99125f3cc6633c8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-924e47440462fc5473ed2cef","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.873","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"098ced36fc7cf0311ab8fd389682438ae3e90aa9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":624,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":624,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f6d8e2b7aee715c42e52","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dbe400a5e736a6e85884f2a5","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"83.099","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8c3bd50baf60eade2f6affc04d4dd4dd178c34a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":578,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":578,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":83.099,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3521bfae022705150c48","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-35b926dfd74c3165bcebf8db","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"76.056","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5da4d4be873db1f8df850dc40908d649da17a365","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":601,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":601,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.056,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6204456af7f0c5e81f0e","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-625ac7066eaba0c884beb5e8","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.606","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"12139dab4df0f0d4ffbf40c5a05e1a57d53b87ec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":394,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":394,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.606,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d76694cc23c9ce524a0a","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6d7a626168627f97b87e4e8c","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"77.465","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c93d951ac21b1fa12a2f32e3fed16f6dda6e3f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":946,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":946,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bf1c6618fdda5fd61ff7","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-29d48ea7e32ea624b3e816e3","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.873","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4365104fa127409a7be0cb6a8828f0aba2faca45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":440,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":440,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-319a9e05fdeb24500a15","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9a79686cfd952f8e225c3c5e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"71.831","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"092c3896d21044e82490a6891416139e130cc2f6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":463,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":463,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.831,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-257a0bdfd71e62f93d0e","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d71db6630ae01022b3c0a8ad","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f5c24891c8dded3d1e377580ca48584b9cbee47","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":693,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":693,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aa41726a5fdc6dd558c6","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5cce9f8098aea0789f5d1378","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.014","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c277f65e5d5490ab7d604c964bc193eab1aace04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":486,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":486,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.014,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0f4228407bfcbad9c401","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fe842cb4481bd9e173b44dba","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"71.831","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b3d4914984bbe9ddecb89c84b10b8847f011ad0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":509,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":509,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.831,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d42f566b290f6a1f8203","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8db47f5d36a5513499cd2fbc","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.282","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5b99e4682ee0b8fab86cc4617eddd0f09164a17","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":532,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":532,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.282,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-86a11ca86493eb29e689","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7dcefd71030ae1ddef8a32b2","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.873","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba6492cd54685edcac24560723a8552d074030a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":555,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":555,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.873,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4b77440522996dbaab55","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9588aac6aa4461cd6f1f90a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"77.465","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd9c8cbc4fcfe00a0041527059e6933745b9e990","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":992,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":992,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.465,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-43843bedafeef6e39733","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2f07c8677914d42734e89685","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.014","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b323ae9793eb8462debe3d72a8db556ab548a94e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1107,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1107,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.014,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-code-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-00c80774f6af3982dff9","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9866eaabc5eb4a5a64393fc3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"code_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"71.831","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1e3486f90ea00a78f200254555890cdc89b1cc93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":854,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=code_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":854,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"code_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.831,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7d60df9e1832dae473e3","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-12e05c19977580730ba1b416","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da4ddce049572e3511d4b172033cabc92fce68c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":648,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":648,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-03d71d6f03cef4a76450","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72c39018e95c5307ad29d713","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8dda5c09eb604726fe67f62d44f03e02365ecb0e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":4,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":4,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-94eb845bbd7bf3d7fa6f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-eaeaba0eb5ef30468322d639","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c7950c58b0bf544d75983afb8df0010956e217e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":27,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":27,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d074342f222b7091d7ac","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d6cd237212763d1964ec43c8","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"678a3b0fcf4cb55e2c4705c7bc53fb180f1d7a4d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":50,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":50,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bbab69fd4f1e3f9dbc8f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-145259d09290f097ea270790","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2abe5a793d2cc2f16d81ebc5738fea62341efe7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":809,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":809,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-10249955cd584ee72e61","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cd4018546cf89f77a6f947a0","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef5015bb82f6f783725fece4b9a08e812277841f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":786,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":786,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-73d03ea846c16c648949","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-885a8881358c52eb076c34e5","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adfa6906144ea0c41e2b5436cd57f34e475b3fa8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":73,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":73,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-645dbf246b5aae4abb65","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cf2c9cd2cad42c93e28cf5ef","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"95.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8d9a8b450a8a599234fff07e0c630105ceb05a46","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":96,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":96,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-24a962279e0cedd647d3","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-89f3b95951ea2d311384496c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"89.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1ae36db428130c57a23c9bfdf7dbda06dd5108ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":119,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":119,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bc9926b96182f45d1a5e","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-26a68b25f7c3cde310c66def","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"97.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1ae36db428130c57a23c9bfdf7dbda06dd5108ab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":832,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":832,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e783f1f4be9973a5150a","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-257fcd954cece177c8095a8f","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"99.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbe91bd546d312919a247da74eb1f44b7dc49b75","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1039,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1039,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ab75e2ad06097dd8a9a6","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0c0e1bc4194683dc51130bea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef8b04f88e87f155d3f467a3eee2a9ff9a9e2cec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":142,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":142,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c23b169d1b29ad4a1fde","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-842e495f9db24b51bab12e84","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a55cc9826d5aa232cabb0109c987c3a7307d7867","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":924,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":924,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3d99ff8ace6a495336a0","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-01ed1dbfe1694ad76acafee2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f400b275534a6ad8f5ebd5a3cda777a6179f2f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":165,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":165,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4c77f68fd0869c4f89d5","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-37d45cdeb766dcdb2ad6f6dd","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b6bc21e8d5e61ee1fa2e010a1fb170a6fc1be193","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":188,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":188,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-45ca744dc36e2f84cc02","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ccc18475a9ae615756a834fb","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"92.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5297f528a780985a158ea7e737e0be83e5490804","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":740,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":740,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-313fe09400c0ffd4e270","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a26ec945019aeef834cb75e","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07a1ce207295aa2851ec216bdb54e36e9ee4efd2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":763,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":763,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1356f2905cffdbba7ad8","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c8523be0ec2b32fca1c32a5b","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"344024397b1f60eedb95d1f5b881c652d53f6c31","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":970,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":970,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e2ca26cf85b218ab2bb7","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6013f5dc476105e31eb1446c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"afb241ac471c627edbbaeb0dc4c416970e7ec00e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":211,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":211,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f20951bcafe4d6b5a1fc","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d218f1ccb94508564510495e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"95.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"184bc9ab3393dd5b91e302c6c10d1f6e41fd45bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1062,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1062,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b9f49c8b7ae17fafa71a","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ff9c86c33acee7c75cabc307","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"93.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"928560471f091a3a058279a947562f5e0f156abd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1085,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1085,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9a02fd6ad33b4621f0d2","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-88f9ffe84bd7c545da2a513d","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"beba9d3766a0ff3f5f98a00898752704af795cbc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":234,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":234,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-68062c5683e83c5d47d1","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a24ac167f6b655a101996836","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92ec45166c93f96fc3174fe36b9e336afc7dbfdc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":257,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":257,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c159f602f9ed1f32541c","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1cc886b92fc4345a96c441cc","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.98","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e00c849cdcb350ef8445c94f4de3e63c301a03c5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":280,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":280,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.98,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d8b0864ef071905f195d","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9dd14a3665da167d2a1fa1f7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.388","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b20cecb619a78134a83a576068d6083b4af7c7c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":303,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":303,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.388,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-521f08e362656b429b96","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-058e944f1d1f212e47106be1","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1d6e5d46b050dd9f3fb4ac38313466b6e88aaab7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":326,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":326,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-77cf539e1e73301f13b0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-66a9211249527f45fb4c4d04","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b71b31955cff8549f409c7c3b495fe1ebaddf1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":349,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":349,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d4d97f30d43ce81c8b42","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d6639a45411d5fc22f8a4edf","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c719f60beffcd6042cd2563fc92c52973d700e98","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":625,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":625,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0181d093f7646021aade","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-95df62c9b337dcc7502f70ab","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"98fa47d5322c1c518a515fd77516cd65c06dc6c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":579,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":579,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-05794cdd1b22da2bbdde","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-91549a10d3fbc989dd8b6087","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"93.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"afc6de624395b4e4d9a36cd75626420bb92c29ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":602,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":602,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e9a449f88db0f9210928","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c22c219f21db6713021b0b1b","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74bfa433f5463117a87f23048b942f0c721da0bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":395,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":395,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cbef40bdf3db4808e74b","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1e03573c0dcd744b63f5e3a0","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e1f264a086ad7b407e2cdcf8d4f4220c6fa5b09e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":947,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":947,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d3e80c440a74dd422cd7","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a3db9b2dab36b636e7d973d","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"89.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a0e651d00470ff8ea215d9ad58a444fe991ba3d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":441,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":441,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a07fce12baeec43f8b9a","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d44e37ea2618d68ac3819225","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85a1a433249527edf51e121350818df19799c92f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":464,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":464,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e13d771d5b552f87d4c1","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9c64e66fe410ee9e4f39a6fd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8fd9fe9aab3d463103aa47bc3c0322addd533537","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":694,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":694,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-144bdbc5be0d41a20fa9","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9439e8d1e6fa9f7a796d702f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5aa1b8a23a76014edba7e1f15a35d19e54b0269","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":487,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":487,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-51e0a0bc627402e5ce27","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e4fa360fecbd575714ac1a76","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"77.167","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9a4c5d4c2878aa9f7d0a130034e2971d41d0229","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":510,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":510,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b6e298a3fdbecd2aa305","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-64056109e47de044124a9d44","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"372f64c9ad260477e90f51c83bab59cc8e40bb8a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":533,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":533,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a7d8efedc30fc5f3b96c","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-73de93c6e0b5126b7b45ddd3","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"88e6a5ea57cc92b8324b2572feb89fef5b8fc4b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":556,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":556,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e647b5e410acd882fd15","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7149a9cf38042cde1c778e19","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"90.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5090559748b88dc5e678199e9a9f1958e5e3cd95","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":993,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":993,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c6ec1d565d3e9daba12c","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a60cb5b6202db1fba820a893","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"90.167","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3115f4ced40c133a19598181b9037d0d3958df4a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1108,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1108,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-connections","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17bb0ca3dc106285c8f7","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2ccfc870184f250df397cb3b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"connections"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fff0f2b8083e71c51b35b4388a26e0f6c6c401b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":855,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=connections","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":855,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"connections"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-53e46cdfbc3a05d73f3a","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5833ef05a18e15944270240b","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.36","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"85c9a59321e0271515de000c52c44f957f7580b5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":649,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":649,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.36,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f41dcb61244f2ec68213","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e0a5146d115fdc2aad5fbe36","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"79.363","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fe33e07f2280b5689bdaa01b727b9ea1fc9a4bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":5,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":5,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.363,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4e5af43f72f638d43869","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d3bfd258d056ea9ddd834508","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.467","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce85e359d83a99d294d0c536468e8d456297b377","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":28,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":28,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.467,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4a18a55213387a8b95a9","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0317fd1fbfb5eace9a30e147","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"89.521","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f0820c5136f2da1c63d0a1e572af440264d07c6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":51,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":51,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.521,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bc1ec05329e10e97ee8c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b0515e6b35fabe54f7873502","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"52.905","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3af87e47474c772fd44c0e8faa831e95483caa05","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":810,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":810,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.905,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-41a5e0755a2b028fe366","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-542b49e5969a00e75d4c4bfc","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"77.571","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90e2c5657cd8cc6a2e2e2625241a8be33d5ecdbb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":787,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":787,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.571,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-25f20d6b246df5bf125d","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72db7005c5affdaeefce2c48","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.683","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070fd9117636d166f5ac08fbe31269ad24409dc9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":74,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":74,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8b05e06a733bc5048ffa","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ffd0c5cb0077203341a1b25e","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"74.741","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8134701df719c0ce51ca741f3d5cb4091dcbbc3a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":97,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":97,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.741,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ed36d714eafc17c66ecb","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b0f61b086829d0041e1b440e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"59.915","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60d51ee1cfc5ad3836fcb1cd11b9516553d635bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":120,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":120,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.915,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ebab2034bdf3f79a4aee","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0c61aa261f571b8edc4e99be","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"89.444","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60d51ee1cfc5ad3836fcb1cd11b9516553d635bc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":833,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":833,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.444,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4b9b473f1e5aa52e6f06","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-68bc6669e2a22377e0dd94fc","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.045","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae80c5528e19a20a9c77242077673140f24ee558","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1040,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1040,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.045,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-20ad2fe88c018f7dbc34","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-eac16ef38c363575a3cde7d7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.421","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b03ad2bbe512ed83d396f1e3b5936565dacc22e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":143,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":143,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.421,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eaad77a76321f7a86fe3","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5f8b63b85a27a78ce42afa62","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.229","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05f1ddbf4a6af344b91d514394fceb2cfcad7f84","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":925,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":925,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.229,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b74e84707ca88087f365","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3a1b676ef389c6af2852d03b","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"85.239","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1b806f1aa42eb7f37f081ecff0e6eba84b4ae083","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":166,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":166,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d0a9db5f038e2f6b526c","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6ab75c57e31dcfa4a6856501","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"47.994","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b27c9f010eb7de259d762730268baaa1cf83a4ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":189,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":189,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.994,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5d561c84c16aabbd8d21","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6aa2d8e1764a0032aa7634ee","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"21.095","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9fa13001765e497421fd39bfeaacc3af1519703c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":741,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":741,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":21.095,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e6913b19b0b947f9bd66","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ee8131625035cc1e73ef3911","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"43.824","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f255e0243a9930218cc5d64a6e14cd6bc63af9c2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":764,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":764,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.824,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2069767e3029e37aecfa","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-be2fd635e49295ab662a32e0","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.373","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52384bbc4b1c4a53c96c68f0afeaa41a23a3cb04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":971,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":971,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.373,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d1211a536a3e91e20378","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d7f0664ff691c4ad1146cc62","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"79.276","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a12dbe014d744ff7ae6dc4e39d35104a58a7a1b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":212,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":212,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.276,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e04bf9a0d0dbb0d68188","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4a4f5db8332ee41d3bca2e46","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"72.309","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2624b64a75409106f003924f9b39759bb6d8420a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1063,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1063,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.309,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ebddf3e5337c7ecea5ba","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ffcafaae384a89f4d013171a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"84.298","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49fc1815f25ece82c7e3eecd73766c626a9db199","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1086,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1086,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.298,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-abd2460f13b2b6c2acab","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f5120bab9e78d473cc66c986","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"88.913","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69519f596aeb68b976dbc5ff10aa6b3d68afe592","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":235,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":235,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.913,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-14aa9fbc20f0dec20ade","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-80ed74102aca8c57a25dca8f","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"89.016","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84edf05bfa14779761eb028695d2b5ac18f3e458","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":258,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":258,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.016,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ff28b52edc32a3443ccb","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7f95f1a83d76cfc425ee953c","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"62.972","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d973e0477e12991f44ca8c883594647618bd717","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":281,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":281,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.972,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-39b25e6ac73014d3b17d","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-84069dbbe777219de7bdc251","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"54.301","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e36ce94d24dddfd421aa45253d5807544d33b426","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":304,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":304,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.301,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b9c1a2b6543f69d7f07e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5dd77a964709446ad5af0978","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.19","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"97e9d9ed5f93fbcafe33adb76572683521edf62d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":327,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":327,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.19,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-467e1af65ee3545ad20f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4f889092847828d94e8518b1","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.823","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b6b72f504766e4cd396e2f9bba6fb9943e7d1d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":350,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":350,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.823,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0437a7f8b0806e4dfdb6","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-50a3fe6734232834f95b49c9","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.504","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28ed1782da72bc0b1b97fd15b69f4329893158e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":626,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":626,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.504,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4f32b026cdad8c2d0c34","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5bf292f1b37c378dab8ff4e6","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"90.214","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a93d9257f80e3ecbbf32556d557210fe176f1131","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":580,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":580,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.214,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-386b7185d042aa9d097c","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d5e9aab5a98f3c3bc2f91e5a","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"87.67","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8460f7e5050b871270adfcdc42b886fcd468d1af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":603,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":603,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ba5c6fad8b2ad173b4be","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d8089548a100d83fed57307a","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.536","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25859bb9950e3281f7b7f57879a9db38daff361a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":396,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":396,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.536,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-244bdb55bf57bc343527","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4b46034c2a9dbb559b6a952a","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.148","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac83f51a39e702892ffa96e7bdd8c19314d0c073","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":948,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":948,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.148,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ac61f31208f3aa582de9","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f9dbe2c9a72e2a352a3812ea","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"51.268","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af56bcb723bd3b5ba3a9c4e04ba9561b116b3d13","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":442,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":442,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.268,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-af189a8dfe47acc1ff37","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a82ff3b50689cc92e50dee12","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.277","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab4c6c0b16623b0ca4c41429f6a8e6458bdc58b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":465,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":465,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.277,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c30fbe2f10903f790683","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fd33d776f6ae13b7480daf13","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"89.758","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58248f2a6d0eca57864fe3b9cce33785f6b40fcd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":695,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":695,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.758,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6e03174d4cf97284d1e8","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-022c024a2bb5bc0e36d72345","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"84.614","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"870422a53e4be181020a0c408630205534b32efe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":488,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":488,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.614,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2f0c1dd3c2efa56f09a9","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7e26dab0446558e5045f1471","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"70.395","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74af1eaca13cc0927388a82d74bcb297fe1c8444","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":511,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":511,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.395,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-474832fbaf38aa491a69","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-10a6326537fa99e2609aecee","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.042","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"79db64d4cd1f0ad22ab61a3dae308a2747ea89cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":534,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":534,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.042,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f4db2551d54390973d1c","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b1b7745b5248308c22f46a49","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"71.8","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f87d18f06f2f0d235118f86d87c1b20916f2179","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":557,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":557,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-766b1181d7ef26ff514b","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6fae4c85779870e7a5091807","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"84.929","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a668a4a54b56440856ad52a4127c7121f54b50a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":994,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":994,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.929,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2f2074669e048f13646f","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8a0bdaad3e85e8bbf77f640c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.623","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49270492b3fb585d9cbe2dbf9b3b9d92756e480c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1109,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1109,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.623,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-consecutive-events","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bbfbbdfc1866624ff273","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ca5ee63b05f2eef20eaff30e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"consecutive_events"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"87.142","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"33c17305a997830ca692ad143570a67af0a672a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":856,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=consecutive_events","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":856,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"consecutive_events"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.142,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fd0ce3ebafec8fb7de34","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-63002af362b190a651626940","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c1f8db9fdb44729de36aa9434d8127c75017a013","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":650,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":650,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b244caf0cf96f90e568d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a89e078689c4a802eaba7b5c","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"08846eba0d132e42b47eaee6e9ac556a4eba3fae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":6,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":6,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4a4de8385a254820120b","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e559b7acb83025de6ae4b986","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"73.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c5c941a5f35261e788caaa868dbb391526649b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":29,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":29,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-00ed2c6d1ddbcab1550e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-74bce5e449cc3fe94a88b1c7","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"774fd558bcbc6ae95ff7fdc73a8e87fe645d55cf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":52,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":52,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7bb16c6155ef59b0dd00","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e39cbbacf89c9528fe483bce","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"89.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e94a5ef7bbe178f59f5bff045b3644ddc359567","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":811,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":811,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bdfb20ec4a619c8945d8","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-80cde5b324687e68074f63b4","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"97.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00643724222bc86b738488a8ef62acd9a348942f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":788,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":788,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-37f9bdaeb0c0e5ca4863","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e3b142cd623147cb310dd3ca","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"90.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0701fa6131ab69143ca197cbc1f7e378bfff83d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":75,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":75,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-692b896812eecc89cccd","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cf224edbf8688ce57e0817d6","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad42a6b843239d6a5b5a5189df04673cb3afae28","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":98,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":98,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-46724b3b1c5c419f0d88","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9084a7d3694216f78a7bd0e0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"37.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2de4e435a55a969a6132cc3562e8b57ee4a48b84","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":121,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":121,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6ed9683b4ad6b1a24838","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1ee7f50930a7af617aef703b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"65.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2de4e435a55a969a6132cc3562e8b57ee4a48b84","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":834,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":834,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0c647eb8afbf3c3aa045","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e2c0afefb5384cef65c3460f","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"037e3d31f634578583937c63fb9cd9df79980ed4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1041,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1041,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7526ce8b3770d4d20213","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f12fba682a67fc748496670f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fa5b9226e4de046529c6d9f01ba3e16e707fe8a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":144,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":144,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d4da8886ab499a3f3eba","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-46004f126243dfc5998e7a31","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0ee5d3fc176dbb0e34e006a14d24e745625985c4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":926,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":926,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-49a5cda5ddd09fc5cbf1","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a916aea6ab17e8e957736943","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2e81a6791ba4b34f5c6327cedc7a68fa364f0c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":167,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":167,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2bff3b2c9bdc861fca88","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1568ab941efcc735926c7e6d","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ea37950ade484dc9cc1fd1b8409d8f4ba1f7b621","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":190,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":190,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7715ed7e5165fe9c75ae","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-161ea4d55e1b5a1571325ac2","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"26.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"decab38fecbecbad91aa844276a3e3a46aadb152","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":742,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":742,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c31c9882d06764556aed","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d509062abf1201a1cf62178d","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a908c3489a4ac5ef237c9dd1a139609238d0225f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":765,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":765,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5177871c8477f0b5a8c9","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-13753adedfad546b19941fa1","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8037e8c45b48be151999543ee62fa04076b031e4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":972,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":972,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0769b87dc452af763763","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d4ef3779a7b65cc21d6b0a6d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2500a98d5d5a6fe2d7d0a572c2a5ddf4b591164","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":213,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":213,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17e3f4208e1e161775f9","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1f1dfde088b5028a5d2d42ac","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5d22c87eacdf8f2989df1d3a0ce5cd747818aa7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1064,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1064,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e4c0f6f9c441c7134f65","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-436b53fca382bcb9ac36df21","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"45.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"676539fb66536b126abf0a3dd629b3fae9d6cdcf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1087,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1087,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-63a4e85e49a887e2c47e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6894d66809c12f78fc594d8a","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"92.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d106c0b0bbdb7d7d6f152134082ed822709481e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":236,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":236,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1e3fea11987df9b9c349","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5d466283759b92c47be42109","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f1a69013da4c4a5c4309e8953ae311383d6f78b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":259,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":259,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd67ddee551eb528190b","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-114341c68f175827b2f186dd","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"37.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"099304f9067072a9d4c9ae7ff9807dba3bdcd7aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":282,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":282,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-65e9d5558d2e8f3aa27c","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-000370e60e31e0b5f509d119","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.828","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f7626691a7318dd2528671bc68c660849d02101","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":305,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":305,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.828,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dd8e56744d453eb889b0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bbca2dae93f84b122f531fe6","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"93.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d091cac21e0fabf4c491d36565dc41416ebe67a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":328,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":328,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6a2fde454bf4cc976dd9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cf112448793d1b68a1458274","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"99.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f6c312af4b970edc552ef6c101d3ac14e01b41","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":351,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":351,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":99.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2d8693fdb766cf0fdbf4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bfdf0ca84c67d0ab89edcadd","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86081c74e1a637bcad3eb0eebed2aa729ed47564","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":627,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":627,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9ed2e1be3f30ed1ec6fa","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-aae11dda322ca1c651066a1b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeb2ca567c6ba693677e091c81cfc71234ed1750","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":581,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":581,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-357031cdb40ced4eaaca","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4b5e9f6d1e3c238bd1831e9b","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"95.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cca1638f416eddc30ede7097750d48c622193f8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":604,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":604,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b5b2965ab919ba0c035e","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e8c9f7a6eb09ac9e0432ecf4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25e12443bb48fff89b5a2c0b1309fb444380f234","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":397,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":397,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7e89c403c3ff58fb2225","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-02f755e5a073c70fd03d4e77","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"85.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2fa3c6da126aa3237478a7bcb61c3d2c04f1e0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":949,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":949,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2b2ce08ea005638754d6","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c16678c457c862d69b4d1964","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"54.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac7f1d0013e3b4b6495fee80b018802237d59215","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":443,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":443,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fded344b07e1cf502ffb","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1ff9416c345af48072d0946b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"37.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32240b676b763b7c8675e46e06e106f40bf812c9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":466,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":466,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":37.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fd6107fc3748515aa3a6","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4bbaf8b70c63f321a3db5a8c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"54.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bed04f5e607318a5a47168a02fd02e54ab9841b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":696,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":696,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7552e412ef14a3bad5dd","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-080e5c5c42bddfd9bea9ef1e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf101bbceab6580c35eb16a228dd89905efd09c5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":489,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":489,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1267cd2b792aacce6e61","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e7741fd43852fef8908305fb","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"52.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4f08f6dd75658bc6518e319d815a1a99e466520","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":512,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":512,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-29fb7efd419d0785f329","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-933c1be599148b342dc2d2ef","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"62.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e37e4e726442747468e3c4495f88af4bff0f8c05","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":535,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":535,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ea528e2406946c8ac90c","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-285b1e84bdef7ea7dd581ec7","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"824edf05be8ca99b174b83647882cfd16559e192","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":558,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":558,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d6294fabbf9dd66fded7","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8e7085a33a660a867842fb53","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"65.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9861929b3dfe0421702136ac6a9dee9ca4d326fb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":995,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":995,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a568b315bdd2a2c826ce","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0d148df2574d0ef8078228d7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"59.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ed169f5ec7bf713525e1bd5f79c593c35380ac5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1110,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1110,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-integrals-with-game","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4d54def23d82d7398fa4","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-539fe191d4721784ae42619d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"integrals_with_game"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"81.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e0a8748975b799b8699ab6c161756f120aea8d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":857,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=integrals_with_game","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":857,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"integrals_with_game"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8e446c0b3875f465495b","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-74119fbe0b6d567ba0e243a1","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70742385be4549255ed81e0a63f144fa43dd60b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":651,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":651,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e1cb6b1964cb17a3cd6f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e845c8212f5c8653ce3115a7","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0b710b4413c35873fbbc388991437871eaab343","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":7,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":7,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-205a82ed3fcffa1ee274","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-85fe471dbb4c4b34ed726f59","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"33ec4245948140055373787dd12ed115e9b44dcc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":30,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":30,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f62747943ab5ec9d1b2b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-759be75dff99fa21eed90da1","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a1aa9a0228f1796f07c39ad4eb81571c60ef4bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":53,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":53,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c4af887a432e959a75cd","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c30de0e66e7b16057bd00096","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06054ff8dd84464f62145b3ddbdbdf2818acbcc3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":812,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":812,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e76da3f24d69b7bbffa7","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-76687d67311239182658d63b","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"77.273","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4cc2e5d53ba8ec52ff74e7e1751f91571de74545","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":789,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":789,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.273,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-84d63037cbea80a6b2e7","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-958f2b36ab4ba582c0dff63b","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"54.545","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b8e9f0350bbf8a0a1c2f6b47deef19ad2914d6a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":76,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":76,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d678631ebb6281def56d","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-366c3d758d92ce2209657f7e","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fc9e6917322dc8578d4a616b88d3713b8dfd252","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":99,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":99,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6da2ed1512da87fd74d5","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-91379d501d7f572be161473e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"54.545","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efb7e4cff0eb3ec6f3c8eb9c5fea91a01d06b304","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":122,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":122,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8b523622be8319525aa9","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-27318c712dbbf75928824f39","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efb7e4cff0eb3ec6f3c8eb9c5fea91a01d06b304","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":835,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":835,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f9c685fbe6cab920c29a","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6484785309b58575801a9fbe","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"665e095eb6f8760b185635dd3cb2ecb5a34583fe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1042,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1042,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-477b49a977dd36de2cf8","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-db0758a5c36104ad0cb11f64","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"54.545","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"86ba16e990b162b516f0bf79795b6572514809fc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":145,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":145,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-40fdd5c9a54432e02237","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a6e405ca114163dfdd818a42","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e9266c0210772e492dc9e3490e54db6e0ad231f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":927,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":927,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c0ffef7864faf260c7fe","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7dbaf6b95af25ee92b7fc230","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4e2dae14e4756b474ef837e7dbca8b2170b2106b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":168,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":168,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7d242f0684c4945e3088","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f73d9be760bce8a65b1d2acf","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d65fe359b33c985ac55d14429cbddbfdd799066a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":191,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":191,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6c924d2842e83071b64e","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8d71a67dc57ee2ed6d0f0abc","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df192e2fa746ca122066ab439e0663b3ddc2dae5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":743,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":743,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1d3681494035886d81d3","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cd4c395b6f7f04642fdc6627","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc2a496e9849e2db8d2c53e122723c43c061d18f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":766,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":766,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1bca6f907e48c2cc0769","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dfda25ef187dc44ae486cdc8","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"250c9d84c7f95d26c3a4b3a7767e4a17770110b5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":973,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":973,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3994f12c48e8fc018884","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-76997c1507ea321921d630d6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"67667259c327273f328bd4bb43dab7609cd79c6e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":214,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":214,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6c503bb41caceed47f70","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d38c1016c80ee1a9fcd10b6a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"72.727","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f5384a64ed33f3c8aca5c3ffee58631dfeed6b1f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1065,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1065,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b29527144d1ef3c0d9f0","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0ac37e37fe25baec6c0465f7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05aeb2be6bc9847946582d9e8cffb67138415a29","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1088,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1088,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e86a4ebcd81bd007f4c3","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8114a2f0dbeefd5d9f2f43de","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cff9e56bebfa904c8f855a0e5a9ec7751a4433ae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":237,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":237,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f3a6ec945567c4a7cf6b","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fa6656aa8c48c4616964f31c","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14caf59a74252e9febc335d63d7e8ade5ac6bef1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":260,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":260,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d98dba68a7fa9d0a63b6","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-528c63dccaf5b001f7f064d7","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42b9482e558bd54dcff328122720ade41583cd47","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":283,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":283,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eb8285bef054338989a0","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4fce12fcf92b4f3a00c3f91f","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"791c13db7ff54b7b270f67ca51f72e4e0cf7485e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":306,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":306,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-63afed4175e39a56b5b0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-802eadd421a313155ab97ae2","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"800a350959ba4e5f45c41b565324046315e831bb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":329,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":329,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6b8c64057711f93987d0","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3a856e4739137e02b07d607d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0663935f3bac70d78991f2989e0755c8b3dd9534","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":352,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":352,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b44c953db499d84c91ed","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d21d2d60b20f1997806caf7e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1cb8292c46cdd0dc1875fd0704445affd67a5086","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":628,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":628,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-76ad86f4f5d2db636121","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b7aad5224461ecaf4ef053dc","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcbc8085dca78ce160150e1a95838ad5cfaacd17","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":582,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":582,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-353a71910860634b830d","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0e0a50f7f29baa842678b2cb","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bb5d180bb35fa5c83904652543c8939e46e1bb38","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":605,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":605,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-89bbe110d99386abe4fb","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d0f9e2f87476e23e74eeab1e","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"72.727","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8fb19c000ebccd5201e5176baeac6e837a827a77","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":398,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":398,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-87db0927f4fd082838fa","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-87b5336ae631aa591a60955f","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"72.727","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c61f9ff3acfcf627652dc358d8a2e18b0dabedd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":950,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":950,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-faf3d9d624a43271eddf","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-52d0ea5a23265f66f37d5698","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7d04cd4b41a5897cd9dbecf570054c9fcc79c70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":444,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":444,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2a0b6de429559b52a06b","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-55ad80235f7d4ef932d15d1e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a27cdd4537807902c09702505ffd273bfd4e7fe9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":467,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":467,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-85ed697daa546f287762","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7db48a25938b24903d507c2f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1228b689cd7a86a0513d7c868599d28a2690e0f6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":697,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":697,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a8418e9bfd18b2b7a1c8","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d2e48c7fa64fc410cf151c01","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"63.636","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6e88e261e10912c86d01e502d282aa9642743c95","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":490,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":490,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.636,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aad79be904666aaa1def","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-52d7202fde255427fa5d2cc8","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"54.545","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3bfbc8734c92a97eb6d253fc1633e587f61fb01","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":513,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":513,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.545,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8f88526f02e985cbe475","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-db1ff2ae66697e192e13b6e7","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaf7e096229a3a7fd9015d37ccc87e15b44aa8f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":536,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":536,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-db1a56d9e54dce820a87","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2e34c3ffe92e5bc888a923e2","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5436f1a6ce1712969e90f46e2165b5e44c2bf24a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":559,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":559,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b991250c112b4c54de14","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-07936b591078df501225b109","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"59.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f72db30c70d074ed50fb4e29cb74b10b2d475951","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":996,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":996,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1b270174ea5a205819da","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a86cf411b509ebbcf28d9048","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"68.182","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a18b494aa752a9c26c18c3a0aba3b64a59e95dd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1111,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1111,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.182,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-javascript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8074359e8126ec94f4e5","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c55eb0f68f89af9abd753323","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"javascript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"77.273","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55c6f375fa09fe5b0b461f21c192c894184904a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":858,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=javascript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":858,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"javascript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.273,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-99abb38307069d363957","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-99a045be3fe60ec1eb8a5458","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fbbb0e40cc6cfcac07119baf5229b49defa9a94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":652,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":652,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fd191374acd0c7e35f42","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-53178750dc748958209d6aa2","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8316b8c0953ad983af59b2af96c18069a036cb0a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":8,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":8,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b60dc3971e5774bb9900","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3dfce42db4d16f4643141c98","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be3cdde9aae4ab6de5b4f9f1b65faaf303940542","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":31,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":31,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1f0b33909641e142f609","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6ce506005f448c56c1c67e44","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"204e4cfd25089b67a7187e79d00c88f0fb679ea8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":54,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":54,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e7ed23097c701b8a1ade","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d2ff6b8f2d5592e43e2e237c","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e3762d3e93fba7870912add9ead5cf965b54e1e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":813,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":813,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-94be3279842b1670cd6f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-08e8e97efdeb064d4a0ea54c","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa0321f015171b52cbdcfdeca3441a031b806d8b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":790,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":790,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-65e2aca9946436a04fa2","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-747af1e9868b2ffbd1c5dc5a","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac4a0ab5a4abff5b9ff3fc239611259d68ffc3a6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":77,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":77,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6f835b907866dfa25da8","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5ec7a670f99f8aa6d21aa51b","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c2e112d356f549fcade182777462ebbbf7d2ac6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":100,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":100,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-05907c7640b9c23cc38b","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d97f062f694a9893d7a90a4b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"64.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ed12082b1e818bbbe861e406728be1d3dc0c85a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":123,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":123,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f970f13eeecee50aae9c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bb5fce4c1211b2acb5072c92","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ed12082b1e818bbbe861e406728be1d3dc0c85a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":836,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":836,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9690ecc41f466aaec307","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ff2e96fd244c09ddb77e1860","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1dcb508fc8fd0af688cd982338bfff92cc1f4a42","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1043,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1043,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5bdbe9051014470b69b4","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e1da478487de230369bc5e8a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"64.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92a97573c5360c55649abc471e0a369a7659e726","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":146,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":146,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-225e293b66bc705375d4","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9d97f4f943c2479148925395","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e72dcfc138b1bd18fbbf9adc6e06c88fc4609e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":928,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":928,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dc39b189309c93f5b12f","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-81cb1c967328fae9bf2d9a8d","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"72.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4f3702c554f5cccfd079cae732e8d17ea25be82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":169,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":169,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-76cd3c0bbb8bb92c8bcb","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-941f4c72bd022ada14c13562","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9150b66ebf98966d1c40e247801627c9198eccbc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":192,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":192,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-af3e809c9b8c6d3c47aa","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-318a65b11067e530892944bd","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e555255a92ca8c5f63f726cacea4ff91e1edcf4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":744,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":744,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-37c1b51c636eb96697ed","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-65d65008e590a3bc2d271a02","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2958932e75daa2f8ce3d9febf5dbad0912ff52f8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":767,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":767,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2e149503aebba8eb1dd0","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-97993fe56cba8efa0b3b1f93","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b24abb09d0ee0af99cc51821b79d8da020a0c0e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":974,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":974,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f100c966786038d43f7f","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-487c7f48c6e81420ea3d268a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8959290f7369214bb9103e90a1b3793df0020ea1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":215,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":215,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-50eaf5b6e6290a279353","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9b6c90b7e6a0b8025ab990e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a545ae2123b5928575371240c86aa761302d5b0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1066,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1066,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f1aa80cfc96dd087e16b","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d49f101c2d7892c11a5387c6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"62.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adde70d7de62c314f8cf2c5ccb6427c3a06f89b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1089,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1089,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d4ba925e3ca63694efcb","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8772fe657adde42f55e78e26","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d467c911be2616826fe18f3674ab5276fb46101d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":238,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":238,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b8884bb5b86c5b5f93f5","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f84513c255f70fced9d77bcb","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ee9ec54a56def5319fde4cac4a3645c6a7227153","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":261,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":261,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6a8d946e7d237416738b","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-27bd83077cc1d43d5545d2c0","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dcb14d56df8e29a372d1decf13c52133f3b8decf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":284,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":284,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6823625276853b75b503","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-139656b40593fc06dbc51b6f","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"64.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b9b4e6e16013eee5f48ddd76ebc2dcc9d093c0e3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":307,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":307,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b1b8aa202c235289ed36","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-074d222b7fa5822cb07012f9","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87a791646b2b24464de10ac11e0cfbf9b27fae39","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":330,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":330,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5917d7f347a94f9c87e9","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1188b35b7832bd26df12581f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f11ebbd2d09844a7d7a5479078976c190853b00a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":353,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":353,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-36231744a1fcd29ae2d8","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-57337f3151cad26f65fab49e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f62ee8c86425acdc3a62058b1104281e7519a3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":629,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":629,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-343de72a4cfe82f24113","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0279d874ea2a6160cdffb9a8","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"733a3075e75104207a01a1fc0d548a5dafca46c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":583,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":583,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-77ad1077409d6484cde2","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-28199ec0502c6ddbcb69302a","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f7fa0820480b5a5735d791e4879f0e0213d60a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":606,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":606,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b4eea803d51d1f41c742","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-077970c01d4beb672163158a","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"72.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7988ba06e98987be2258f580a5f528fe4cc8aaf8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":399,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":399,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-275e5bedcfda7f824113","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2ac5d6fd0d6ab9d2eb27d022","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ab78d3441a34bc6ca3bb3e16712e7b163b1f9fd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":951,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":951,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-549741fc3c24e182b0c3","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-769e0f5da34f372b6b85efc0","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4a97b34d091bc50d08f22f3fbf16fcb00c41876","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":445,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":445,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a260e83a2939e200091c","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5f9a98dc2bee8ba4c1b8d54a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c6c61e9abc8d4956219b7f30b5364b6de1eccb33","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":468,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":468,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eeb8e12157815f12cd45","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f4ea8a308a8081a846df4af4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce7b26fa59053060858c70e5a68f598b16e2ffed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":698,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":698,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5674807fd73d54588236","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bf547ff02c09e22b6bfa7ca2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"66.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"004387b08a4926e081b953480c9d5599745920bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":491,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":491,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2b33eccb7e14eb9074c1","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ebed4d52cc7537406ef2839a","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"62.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ff3b4c8c09f81b381e1a80ed77b359ecf1291e1b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":514,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":514,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-729405dced13c2be5403","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3d447cfd1227210396f9aed8","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6e46bfef307d9b850f9ff8c10ee2d79471678f48","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":537,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":537,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-af52429503dca7b58236","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-40799742423263cbe381a282","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bcca6418de620681b731ac4bfaf17e2f1b1cba6f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":560,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":560,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-abfe31031bb92a76bd46","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f992ecdcd6ec1bf24029a65b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7771b09f8d76374eedd7d78e0d3c5554256ac9dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":997,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":997,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6248fb4debfa44ec200e","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8150a47e7e124ef0ba3a05bd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5633b3efbd9683e146395b59a30afb27ff8d0e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1112,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1112,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-logic-with-navigation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1bb2db4e1d4fbb60f72b","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2331f3c762f47564409e5e3d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"logic_with_navigation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ecfc88d2a768cd8b4af82a7f89c5159326aaa87","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":859,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=logic_with_navigation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":859,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"logic_with_navigation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2ce3c822b24053c6011b","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-03ce52455d0732faad70a492","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69c2c652d9e2356c2f0fcf040afb44820036960e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":653,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":653,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0b2073fff1adb312f73f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-abeb82f6431083c45816613b","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53365f0707e58de5da4f4a60071b5c01c8e54335","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":9,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":9,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9b9c8ef2906f93fed7a1","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1bc43abf4a0c4423d26984fe","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9416dcd733b77dbe6144e3dae3c099ecda5b02b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":32,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":32,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-110882cf971303b5ecd6","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9ebdc994ba29a1e451440e75","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"98e7c5d26555075b9135de770cad211407997371","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":55,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":55,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-82318624268f602c90c0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-90c760d7c80d3e11613a815e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b703cac08ed9f58b96943e3f9cf6528c1f339ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":814,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":814,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-73f662ac2ed8f290c3cc","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d80526f1a0f02e9382c27344","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc63cc7ab5c27cc85ec210fc20a9d53e99e471f8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":791,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":791,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6fed71aa5293514a7b16","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3872a55d077bd4462719d709","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d40509563d7c556e5c0186560550bfc3b47a501a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":78,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":78,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-67ee0146aa8bab768930","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1d369baeb4bbcb5c19bfbbcd","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4920a284a0eab172a673ab97fafd2e493da3fe65","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":101,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":101,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-634a4c8680d621fbaa52","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e7a3bef152259c8a5a5a8b92","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"97.059","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d7809d86c29bf444a56ba4b1aadbd675f8e0bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":124,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":124,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.059,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1408e27ed928b3f40f2c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cdf0ef985494fc79264e8663","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03d7809d86c29bf444a56ba4b1aadbd675f8e0bc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":837,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":837,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-452be7a05673842c0e63","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6487f6d1cd39c1e3e353109b","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.059","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ecfa91690263cbc78fbd1dd0bf17daa024e2f19","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1044,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1044,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.059,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7bbc0254754b72a0ff93","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6f2310373616219f579e047e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"12ab0cc08413787814e87922a50db60b7bb58922","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":147,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":147,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d52b3ab1e23aec33a306","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-db8411c70000abb077a13f34","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.059","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62ae0248c4babbd784df092bc163ea39bb988f55","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":929,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":929,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.059,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aca6fae93dfc8a13afac","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d9bb91ff42b86bb33afe61a2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1939113768d6197ce1f821fa2eecd2f7500136d0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":170,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":170,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c2fb14c3386d9cf09c43","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b4c8f4f551193268cf8e12ee","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c86b447f4da3f87affbb49747aff88de3043d6d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":193,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":193,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-afd5a5e76b4aa644d2fe","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-265f5d9081d51cf17cde06ab","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"92.157","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"569495e4de20d71cd1c2f6a981b29fd12af2e92a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":745,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":745,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d6c9afd30ba428a4567b","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-021a0b8539a7d4fcd0385548","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d4c82b116c4305ca28768323e7ef7b8adeb1e518","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":768,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":768,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a6a04df2afb05684a230","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5cb2779bf865a5e0b2707a8e","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13fa95031004b4fe4cca73bf1a0abcc788668fb2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":975,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":975,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eebb9537409a137a8694","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-189c6fb20cfca8140d8054f9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f30b92933092ed77b8a922695340e153d4b5b104","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":216,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":216,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17279b976b030c131058","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1723f21aad9c3dcbb97b242b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e72ac34f877a780ec2de0bad42dc81060675309f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1067,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1067,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5833e91371b49e6e3af2","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-229da68ef9c1649ed24a44d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"93.137","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fed78320ab68a65fa5432301f6e69bb82b1d669f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1090,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1090,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3b6846e0f75996f20ff6","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5e264d28e4d78c0a63511e49","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a71da476124dc5f13a29dfa3946fee3b589a9dd9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":239,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":239,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a4ed4380c709961b1755","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-80c78b3777b07896cafcd6a8","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"611eb2f64808781586c9f53b27ffc950c3813fdd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":262,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":262,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a0cb061604c579b99393","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-09655dac0c85093c2a3b045d","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fe32420409e1441a45f5e8815c39ce287b248a03","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":285,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":285,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1959ed346e75fc4cc111","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3200fe37c88dd084da0b589b","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7809aae15f4ca717afbbc7c4bc26a655894c9067","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":308,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":308,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f56ad275de1b72a93b78","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-11c1499d3bd93ac50d82df74","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb3517998444d0874f571f1c9afd688add6a6ec5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":331,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":331,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-600fd1f5c3e005b9a3b8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9ee73aeec0f783307ac19fe0","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a220a827aa8b546bececf539de32288133766a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":354,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":354,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bfb61ce039a0f8829b8d","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9b1ad40457ef0d30d40ff4f0","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.157","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bc81ccaf9dc6ff5c597a9f31c96c200c0afe0369","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":630,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":630,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3d82c5d086c7e4676a2f","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-010648d5e691a7df4c7de692","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8973aaed03586a50e760f805ab38f5f632cf1ae2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":584,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":584,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-654bde577c326a04b3c4","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9be0a4541d4f9652a3e062f9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e80ea6ec4939918620af9c772e46c71e9aa3bf6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":607,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":607,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dfe62b6d7b9a72284860","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-acb70937e09be2b80663ec07","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.059","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b8496c0f67206e60c17061e49c2b142efe436ec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":400,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":400,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.059,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4335dbc1b08ff7826341","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a3001a8a5cbbfa3d6846db1e","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.059","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c98fc259ba7bcd3daed347dc118030fcfb9b991","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":952,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":952,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.059,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4a5c44fc2cb090578ea4","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-157e4fad37d76d542167dff1","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"951e8be02d0cd0802624ee3a9b0bd98ca4f1d729","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":446,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":446,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aa198f549dbfe4a22d4d","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6c4786f2b419296f08cf24ca","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0679d7023cb2011896bf9b08f7d88594b7b5ca75","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":469,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":469,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-76ba902297c7b6e3b92a","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-238971b84d4cbfa526139703","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc8356f8200532d1f4ef094750ecfb91b1095371","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":699,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":699,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eb9811bd9dac224ceff8","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f5e4b5cde6aa5a9d41eb1d97","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"91.176","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74ff99d5ad99efacd65827317adfaf395e5ced3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":492,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":492,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.176,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-38a772f069b6272f062c","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1b378a919fc8a446c72e2ce6","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"92.157","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6e920d164db214c2f3fda5b119542c558504969","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":515,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":515,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-15802dbebde64d7e1ac2","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-73219947fc6ebdb60a56ba90","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"93.137","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84994a5cdd99555492b38f90664d7b88650f5768","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":538,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":538,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.137,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-319b6ab12aac483edab2","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7059f2c6f7878edb8810af9c","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.059","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00d4ef79bbc86373d8103a039eee949b7779a77c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":561,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":561,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.059,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-30841105e23a7781b691","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-91d311a1a217de21d6be0831","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"118e953b87278e5c1cfb71d6115b49b3ec10b51b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":998,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":998,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-415b990b4dc661a36bd5","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-193caa2964a5759d53049f34","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8edfe4ed54b0a66c7f9d7d8edafcb2877796c83a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1113,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1113,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-math-comp","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6b803b516492f55f3116","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-987738cc6622893d492a63ef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"math_comp"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"95.098","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b276d7c2f7a72b6a34227b40b171c0bfe3f81662","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":860,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=math_comp","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":860,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"math_comp"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.098,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-67937beef7033912467d","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4fee1a40277404f05fdcad7b","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.845","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6961e3626c99721cf23ea4b5c6672907a5722616","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":654,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":654,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.845,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c03d94da7d92b55e8dd5","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a5a095fa0601af55ba0c4c3d","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"89.458","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"756d5f386e083272d52d4c526f17cb5b3fc9a4ad","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":10,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":10,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.458,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6ceb2adbe587eb3cbf93","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-165f05483c6c1d537b4baffe","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"92.17","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c300e400745e6a208f67723b5f44930e4e5bf831","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":33,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":33,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7e72a18163e3760bdd44","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-40de6f7670b15c33141772b3","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.376","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f22b82f199acae3a65ce0120878b773faee854bb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":56,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":56,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.376,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-62c99e864d02809f26e2","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2479f1e4aaefdd1da28efa5d","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.227","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13ed2a00a26d5c001ac14646da5f8979ef7e58b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":815,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":815,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.227,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a18b813018cde93fad10","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1ce038220f738ad6f57cbd28","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.796","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72c6ecfc085b3a291e306df2160242fd23d25a6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":792,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":792,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.796,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9872ca7b92bb886f10ca","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-33403945336ec7a1c5b7ddaf","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"87.858","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab0519fbe2a2b83b965f171d9b6d46b44d1e1299","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":79,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":79,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.858,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d59648a1f67422613490","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3fbd277127c5f732c256c337","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"90.671","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"91f8edbce1a0380242151723b6de1c5af30d76e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":102,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":102,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.671,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c58a489de5f950c0f9d1","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bf581f39ad71081b1909b225","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"86.531","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"407bd5ef9ab25408225f7a7ac9eb0dcdf64fedd1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":125,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":125,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.531,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5c91be010f5e7a40c47e","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-29540e4320c19b7456f5c3bf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"89.081","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"407bd5ef9ab25408225f7a7ac9eb0dcdf64fedd1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":838,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":838,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.081,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a680b5375c59244b96fe","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-906443d64ddeee123ef62c60","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.168","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0509124d2dfc60dd3a2e53438061363a387176f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1045,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1045,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.168,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a141fe39f540c3353c52","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-13a3f9c0f85af50a1495d534","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"90.624","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cecb3bd7a07e8c46a9901ab7e8727a1a926a2631","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":148,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":148,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.624,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-589aa8dd61a5ca32e5ee","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3aaa812e22f2d84bd03cda1a","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"91.286","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab24bfaf8a27b09985dcfeff45104b4282b76b3d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":930,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":930,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.286,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3a6058e83228cc116e3b","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-70aba71a98342f8d09da80dc","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.102","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6777b74a016e12c651fde5c7dc206513e0d64988","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":171,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":171,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.102,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e233aea0870a11ac348d","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6978336a4a53424865c696cb","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"91.877","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"788d3de53258f6e84b249d42709550d67e650be3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":194,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":194,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.877,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-db1aa3a1b6575f528fcf","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c38ed2aac583cc27650bc61a","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"81.801","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a37323579957611d3cbfa1087ea567f25613101","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":746,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":746,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.801,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-91b78c813e2134d28d67","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0b42abbb6c0f0d22dbf06473","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.492","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ced309ffd1831f86b74166e1264283653bda340","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":769,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":769,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.492,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-db1c1288db1073390e91","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-61f6decb39ec38b44fd26965","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"91.794","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab1815d4de0c1b33b3f1a1daec6de683f6a18bcb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":976,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":976,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f8ad980d2061f22ce894","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-835700392b3ac0c98d4907e3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"89.047","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4aa3ce360bfa4d2a21825591d1e105ba65f304ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":217,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":217,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.047,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ad10cf04e1aa32cbf65e","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-327a89f97a32be720b10d275","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"87.514","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ca4fb4ba49127a8f5acca366a7b5fc53d74290e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1068,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1068,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.514,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9d18c97c9ccbc7db3c9a","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bda96915a01dd19e6410608e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"89.841","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6003577b94b5141467492d8bdd0dbc1caa703069","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1091,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1091,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.841,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aebacf8998e036c450dc","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8dbcf65cc1d97fcbcfc78f10","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"88.566","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ec35a0f426c12b73aae7d37c503f369770fbbd5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":240,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":240,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.566,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0427f0ea6024b3268d57","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9c924566a62069d1c9674da6","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"86.997","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"398716dc34704e71f2dc3759748b8e762b340713","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":263,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":263,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.997,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3e29e1da4bdf7d050fbe","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-684e417618ebe4170bbf4e9f","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"85.73","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"375f71893e73c990ebfb83703643c136d3443dd0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":286,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":286,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-065e1b22be352196595e","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f2565e8821f22a3a83ae0ee7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"87.982","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"09cd78b9693f486cf827dd533a177fbfd0d35df2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":309,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":309,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.982,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4326b3cb75b9c21585f5","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-486d8724bc394389aca0e5ab","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.474","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2773c519d64bf72fb621763d96f575773c3594f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":332,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":332,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.474,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d7255aaa4322a6003621","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-241f3a13b8c9a813b92b811d","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.351","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"72f48f70e9258d9c234ccb68c986ab842bf8bad9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":355,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":355,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.351,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d7173fea2f29ace27105","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9c400eb2f9f2a6a469b3dd69","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.646","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"56a1ccac3223e06242192ff2dd7638454279983c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":631,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":631,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.646,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17a87ce73c340f59e6c0","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7fa87c57fa2bf2146e47e78b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.693","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fbab401c390e9e82566b14194972af76998784b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":585,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":585,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.693,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4feac889e260ee3b3fef","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2f95494b6d3f1aef54cffa13","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"91.535","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f728f6baa9ec345dd5fd2a49f922f295930d988","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":608,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":608,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.535,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-99701413d63793fa119a","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fd88003515ac9ea4abc4937c","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"89.239","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"84188af05c445a50b8f66cdfe4eed5bc62cfe86d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":401,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":401,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":89.239,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fce9c3c85b3183608b7e","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-05f6f62602ec07e7028790f1","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"91.213","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20338faacf41dd70648b74a5c52e40ac1dd65372","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":953,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":953,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.213,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5af01a9836e688d72ed6","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a55f36a7bd439375ad48ebf6","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"90.027","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2db8de888ad97233c8b9c7a96d60f36fdbdcd055","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":447,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":447,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.027,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-84dccb04fec93a7970f8","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4e640800aa002efc63b893d7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"90.299","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"166ecb8360b497354d5739e4b4a74682d59e8ad9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":470,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":470,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.299,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2bbe6d398a7d3195da26","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b882550c1b261b48ca8719bd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"91.649","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"549f29cc87e277c05edd566f263021ac6a3b80d4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":700,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":700,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.649,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3bea9cc79c963e1714db","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a9b70ff396313e27534b70d1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"85.614","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"336b2e5b379c72eb7a16d9eac1b88d92c26ffc19","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":493,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":493,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.614,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-df05dcbdf42e8b6330fc","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9ad5381f1867756531707c04","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4cdee7947fe64533d6630976a8370d4aed55f08","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":516,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":516,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-607d2c3cf2950b0e1d76","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5c54a1eccab9c940b0e86d99","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.762","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fe7093948bbfd587bab25c063b0af507e2ea23e9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":539,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":539,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a05a8df2b04b57913e5d","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-33d6d0b66165304565d933f2","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"86.934","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9ee126e52304bbd8843e694aff0a7ee8981bc1c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":562,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":562,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.934,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6df7bd46e842235e59bb","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6ee2a2e5e0c55ba48f992e61","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"87.733","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9cd3c4bcc4c3f36b24412ea078a22290fc54f88c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":999,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":999,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a71dba6dd0fd1d1dc380","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-86b02ed1295edec9fc934c11","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"91.166","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"942ec6fa383778ee0c5ea1355a25d93e0e5210c2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1114,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1114,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.166,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-olympiad","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5e60102220fa283e464f","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-731b9eb4f239a07b668b5cdb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"olympiad"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"91.151","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b0fcdb398999a1a3fe06fc4a9d509cd78156983c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":861,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=olympiad","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":861,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"olympiad"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":91.151,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5010824d7ba812c240b4","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5c50db015afbc91466afe557","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"77.65","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2cc11b10a272a2f9936995cb6f65f79ddf02abb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":655,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":655,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8589e5f2d55e96e3af1d","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8e965b628834ed88c6625164","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3c8990371372491375729e7b872dfbd8f41a8f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":11,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":11,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ddf25e13585b62e58db0","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e45df10f369da06f82329f0b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"62.817","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c6ddb78e7f20b0063848641c94056517c0e2b373","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":34,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":34,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-20e9f04422a5610ac36c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d61d1bde9d58fa0835c665a7","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"65.767","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1257b65797b921d4739a2ff0e14398f051075363","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":57,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":57,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bb5ed5c516ef42d5cee6","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72c4c962dc9f111ef185fc74","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"67.867","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cf5716f1aa1226b36c252541aaa3828b2210045","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":816,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":816,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.867,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-84e342de659486fc0c5b","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8f35f5850c1127a3338158cf","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"65.733","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66d44ec2d66281777fa4c901e80e77f5901bd3d6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":793,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":793,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-260d710ab059261082b5","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6013e4e0294cc138b516c295","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"59.4","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"354eaf7d5d31f7d0ae089cc66db0380f2d3267e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":80,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":80,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-016b7ac864770295a3fe","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d6131b9a63d8d4ceb237923d","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2a5538cadf0dac3b65316f0eb047ba4f7c2fe7b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":103,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":103,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9e4e2066377991549079","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fa26bb7be71b26e9672bf03f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"62.383","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"623a3f490f7c9ac600ec5f62d94448688bdf1c8f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":126,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":126,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6296783add6d4449a65e","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-332a85fad9eed8fe6d99a85b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"61.7","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"623a3f490f7c9ac600ec5f62d94448688bdf1c8f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":839,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":839,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dd3fa6977f27e26874f1","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3ae9615b9e98804251cb7ee0","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"72.833","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"257caefa899fad1ad2c710f024b329fd0d3e96fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1046,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1046,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7fea89ba6ceb93b39fef","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-eae7cd7acf75af220349c3cc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.217","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"17289b170899c9c8e90b80652ed1d1ec88d39dc4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":149,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":149,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.217,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9273a19cd31739fba7e6","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d505c77198f3e3b43ff7800a","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.633","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"435e1e92bdee7946f3ee1142cce821673455e6ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":931,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":931,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.633,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2b44c453ac2a0cfdf024","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6e4bc1323533e88330a5e30d","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"79.867","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4dc2ff314bf78e21fe5cfcf327bae46cb5fae69a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":172,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":172,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.867,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-831d518e961d9b1e1ed0","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8aa0c6e6301f7272002efb9c","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.517","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e9755a42d99165aab47bf76e891fb32ee4feea82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":195,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":195,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ff6a5938c11e2166f5cd","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-62d1fa0d06453a4fed7191fc","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ad31ce73af0c6febd976047c71d28087feb388d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":747,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":747,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fcc29b3a6e32b14dd2dd","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4f78a8cb59c34dcb7980a3af","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"76f01fe0aa60e3b99aecef63f085726473612298","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":770,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":770,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9c583b26efbea422b5a1","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-43817a084c9a663b3fd50798","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.067","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6612d0581ba67318e127292725e7bb1c3d85a6b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":977,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":977,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fef619af6b2a1ae14de4","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-404d2178f7f02e3240c0bcef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"62.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a7d912f9adaac75b98d5ef075f58c168733e6ed4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":218,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":218,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7a193814da666c8c24f5","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5e9544cc3cf6325bff7b69be","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"64.683","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"795a510d5b3efd3d6de25078db5b12816c4d2fb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1069,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1069,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.683,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dbf69b77542cbad854b2","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f89c65f2b811409cb54d5c61","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"47.983","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc28dec4aff82469d0be2c6a4ab5ff833cf18b86","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1092,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1092,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.983,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2bc9c73725efea9c402b","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f5323499e1f03a04893bbe49","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"64.1","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1bc35bc6e9c446c3f6765a2710c381de4afb571a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":241,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":241,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eb491eb004d34ae14e0e","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cac11ebaf650051a0bcdaddb","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.05","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90decdcdcf6615682baa9e92940650cbb1e5648b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":264,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":264,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-799f5dd82c1e2450bb22","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-88d799e5f82f3ef08182d60f","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"58.883","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d7b310ef51ee0b2aca6d059424652ca31527feb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":287,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":287,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.883,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-275f94bc6d2c0c1a4394","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-824a85c75018a098d760415a","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.649","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d51fbc12cb2b6787510f1952e140caf48165063b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":310,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":310,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.649,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b97198f798e93a2c8b3b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c076549562624dd79c231c1c","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"63.433","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeef5327c299b76c3e8a03d9b17dfc2f3b4ffc62","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":333,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":333,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6b3ca34852f6d4d92f05","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-af761000c28379f88b4b0430","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"71.733","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f4f63d78454653c236b98411042b5a88897b5ede","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":356,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":356,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-520e23e5a02711d840cd","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c9eb720f32bfdad6a43fba81","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"47.817","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40cca04201eed43c8f4a26d416e73799fa4e7c94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":632,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":632,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.817,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-af9121b500f1e8873a38","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a9a53c00bcd6534b1baee17b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"71.517","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"697ab501dbb2f1cc7356f288a981f8ca6b3d21f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":586,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":586,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7d984be94a3d2daeb2a7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-63587fba811d83316052aa80","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"57.033","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ab77209da2e83bc5079a09f2af99e48fd20ecde","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":609,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":609,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":57.033,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-62e85166f654daefd407","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dfb0cb4ad7d3abd9c58f6844","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"69.217","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eab2d9f35fa4d89a801517f4c89245a9d87a7bcd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":402,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":402,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.217,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f3e7f59b684e90bf75f4","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6b8bbcb5e1aaae146e9a3667","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"71.783","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c91a973655b87923eb7771e6aba409093183eefe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":954,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":954,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ef19d147cdca60eecde7","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-672ceb7f221aa584ca01ddb5","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.633","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dcdc6b5b227f0ccfa2e574d864a9722662521a94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":448,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":448,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.633,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d34ac7c6208fe7392889","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-976caacf7e54cc9496be24e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.533","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"630414489ea136733aff1fd7a50f9fb4044d4414","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":471,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":471,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.533,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7fb816ef7d895822ac1f","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6b3aa1fbb72ca5e6f69e146d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.367","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c0f265481d4667355b936c9b2154abd6f628b0f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":701,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":701,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ff0ce8507af5036ef403","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-159666865b621aff942bf9a9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"55.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7bba8959aff8ffd119af3e0caf7926b420697fd1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":494,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":494,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9addd176472e3e2e57d7","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d9ca9253bc60eed1733be817","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"49.433","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b25f991403db629667b38b959b7ff18b5af8b32a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":517,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":517,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-540c74e71e92779718ea","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b90a6c0f35f8bc1b34f33a67","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.617","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e600b5f2a5c33dc910108a4afbdde6481a9c08b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":540,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":540,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-34f9caa736188215c787","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4807dc6b7cf7ec79852d3b5e","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"72.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b16e622cea182998f5cee66528825a03986a58f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":563,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":563,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b42abe112e2ee4948fdb","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0df89b930d7462469ec8f9f8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.9","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c5e5549f6e90ebf6036fed20992e2d351ae97c98","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1000,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1000,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b7102d26e80ba7353791","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d191c0c3dcc74843ee5f01a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"79.467","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b000c4b70af063da56f30774fbaa9d8bfbbb5a0c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1115,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1115,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.467,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-paraphrase","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cd00552e197be921398e","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-30c574db5252a9ebb3693dea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"paraphrase"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0052c85905fc23dd78c5e16eeb80be6199726dd0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":862,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=paraphrase","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":862,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"paraphrase"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-131923ae6c6a1a6910fd","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5a2538f6702000be830fcf81","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.719","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8d839b4c657571da6806969a9717dc51ae30d567","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":656,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":656,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-70d2830c19090c4ba109","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8bdcba592d46fdf5a9cb5ef7","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.45","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a05f0235409df32dcf72051445327af5ca805630","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":12,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":12,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.45,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ba077228c8722de4a2c3","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8608f9e240171fb783e27c9d","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.476","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"07055bee7c673c28b3931893bdccb79e4a39f986","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":35,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":35,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.476,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3e1ec6e2c6a09a234543","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c2b1d214e4926894e17f0175","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.409","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1fb8f6e37f6c7f620f0b307b4cf2b2212c7839c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":58,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":58,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.409,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dcad4deb5a707a7a7bd0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bc9dd299fcf9cefc76e09d1c","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.645","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d99cc87215345e0206b7a8ab75274d44ca4efade","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":817,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":817,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.645,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ebc647cea9399ff691d5","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b61cecd9d5eb45324d9fb768","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"74.731","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf40112a41d417f12ce68d53936330d5f1767d17","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":794,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":794,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.731,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f5336431346ae11620d1","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d1d7234a1309b9de5e756af7","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"56.975","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b30ae34ff19ed9a92e04fbb6dacc2e8fad79ef27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":81,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":81,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.975,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fed9bc130c0f2ae9dced","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0a316831c15b47c74025ff8f","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"53.411","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dfd3630a9b4a1c2d145e3d8174cf5d77b4cb122f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":104,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":104,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":53.411,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-31f6e487d79e4a77013c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4623b5e6645cb2912358d702","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"46.872","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71d05a11f4e7b2c372c140c3bc79d57941e78adc","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":127,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":127,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.872,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8ba378fe2498295bc9cf","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f78a5723705575eccf751aba","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"58.192","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71d05a11f4e7b2c372c140c3bc79d57941e78adc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":840,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":840,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f82ae47990cb9e0f6660","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2734ecf8c0816952b137e812","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"57.746","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a2cfa7798657b5a4bd4fd107763749ae6007c24","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1047,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1047,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":57.746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ddb9314fbce39d6bcb28","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f445ab16e6ec999531d0b47e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.391","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0401cad4638ed97538379f0e0a088bd02446d18c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":150,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":150,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.391,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-82f8be9bcba7e79a38a1","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-84e3ee5945f8d4e5db95f64e","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"64.222","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"430d5c2d2071c9b718d5ef3b2f074f9bbce81811","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":932,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":932,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.222,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cb241651b86e8d599e41","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7847cc0c9ec65a2676d217ad","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"74.128","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"392d8aeb7fcf307a3d3d1f90f1f22326f889c0ca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":173,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":173,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.128,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bf972e511250edbca016","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b100b566a58424dba2e08896","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.753","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e04531830b2c8984808fcc2663c0495b7ee498d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":196,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":196,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.753,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8d166d240d2ab655c9f4","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72bb67aed820629c6fecca72","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"49.461","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1eaf77f07d5225bc0f8228a20ffff32017a1a212","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":748,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":748,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.461,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9f6e5d95e7fb9c4d3d98","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-091b87e81453f5494832cf4a","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.693","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"863ceef60e3a9d9293cf49aff6aa357051a9adc2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":771,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":771,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.693,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17ba767e8bcb4992af9b","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-03c6501f8936517a9813cfee","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.365","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fe73d38fe2a4cbd3a4fb4f08921e2f08a3d9320","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":978,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":978,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.365,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b326f6515076a3652830","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2cbc7f5ce442435a541b04e5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"60.725","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cb37f11a06424e0932fefd1033de6984b15ffd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":219,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":219,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.725,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-44993ef2378992349bdd","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-164d60f197cbd294ad2d8d30","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"60.236","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa0a88778da2c41182e5999563b7ca66d77aa87c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1070,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1070,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.236,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-646f9559a446d8665828","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2a15c1f947efdd7490e4829d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"60.255","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d0ecd2cd5258a14531f4d630c3951fe30a77d84","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1093,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1093,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.255,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b82d6a903be65604d747","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-18e46f17f95518a53a33d3f8","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"58.427","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"33d64d3c97146aa5bf1f9421c58df15352ef1f70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":242,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":242,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.427,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-de61db2558b048a5c754","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4b4eb94386558670d7b1c803","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"56.035","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"488504ead5ec80f2a5e170a462c2a5589294add7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":265,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":265,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.035,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-afcf9fece05d76c13198","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a0f9171f89d37f8ff9ff0ca","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"43.875","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ec98b8b16fb3e209795a84eee8fff72ec326dd3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":288,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":288,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.875,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7f4e851a4a1452b6f060","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-559cc66f8201553c8129faa5","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"39.133","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fbcd0c8da23615c0ceb81a13a6e966c166b0a79a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":311,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":311,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":39.133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1253898489a5c1a49837","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c19a2b0f6ccf772f1d0c0fc4","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"65.901","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"163f96a82313ebc0bcdb2ff0a59a82a5475b47a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":334,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":334,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.901,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5e5d07f76237cc5e5814","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ecfa8c770610ba44a21004ff","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"74.089","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3209022de82dd7ed21344d677a5bb1c5266ec2aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":357,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":357,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.089,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-231dcba4f4db176dff8d","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-de299cc19377b62a9dbd5cb0","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"51.201","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5fc0b82d2319a110d40d06f85458f19b0440108c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":633,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":633,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.201,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5bfc2d25dd437f500f85","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-625e16f4c562df72302af553","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"79.051","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5de26be80b881104eb3970d02efc9348f0adfb11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":587,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":587,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.051,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c8b89b2644c28e173249","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6cd7b80a9d867e7cb08a0968","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"69.678","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c3e2f31508029f5dd4695a7a626554be544ce93d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":610,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":610,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.678,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ec5ee59d71cbc5abb7dc","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-78570f07ec6e774588740382","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.386","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"262a7b4ba0a1ce1e64f948bd47627687455347d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":403,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":403,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.386,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c0c82e10de3b660a156a","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-54ed4e07bda96ce8292448ee","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13dfa07d248cee258bd8c0a9fa980eb1b64071a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":955,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":955,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-596b387bf9225d576e8e","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ce14e3ee060d3e85d28efa60","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"58.091","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"921d8065f20fed8902d6884f5bbac41378bb6cd6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":449,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":449,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.091,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-61910db4509a71d1bf22","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b33d0630dc27f1815b1b4775","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"55.717","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27ab0bb2f4053f4cc421fb9b522be96e84f9874a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":472,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":472,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.717,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a9715db8b6201b9a7428","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ed091f24da25c6ed2294db4c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"72.584","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77139c17974e19d5a237f903bce3687e002da451","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":702,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":702,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.584,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-497b56a84be6b0143db5","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dcb38d2b912edca630ae1232","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.174","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7aea9b874e27637260a10235b6a09dc13afc88e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":495,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":495,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.174,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f5b774f03b17ffd93a65","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d33af1359d68b4cde7fe1bac","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"42.746","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"06cff90dec0085fe7561b3e8e0e5b12a39aa5c71","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":518,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":518,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":42.746,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0a5c0fd963874cdb42a3","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f4b907d9f61e3933047b5136","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"52.468","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54214589ca2d7ff0ea772248a92dc9babdfa591c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":541,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":541,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.468,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b7171b0073cb347616bc","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0ccd139cbfd76b8f7d5bffba","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"58.717","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82867fd2456ff4e7823a30fc1c4e9a7e3b37ba1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":564,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":564,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.717,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d7eec18aaf863c30b407","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ef5fa1164989badc7cc38a17","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.537","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d14a06c40fa60b5fb3f9e5eb968c1c21be3eef0d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1001,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1001,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.537,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6043f60275687906561a","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-86def491612e4ea19c6b11e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"55.762","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d555e32aa480d0a92caf5b6a9516a7715afc9a32","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1116,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1116,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.762,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-plot-unscrambling","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-485789a5fcc619da9e55","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-831c8a70750bf1423e020e6a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"plot_unscrambling"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"58.562","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc8011171b322c8cd17c37720f99875085a5f54b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":863,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=plot_unscrambling","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":863,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"plot_unscrambling"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.562,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8b442c6a8b84c06cf055","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a3089444d215a42ca6f86a71","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"65.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac35c9cc46707fb24441e0ce606c67063d1d77e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":657,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":657,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5ff4be144fb09d32a526","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8a7ab09cb6441bc2cdae92e1","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"40.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"854f991a9bfb49e1d0c7489fa843598ab977a407","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":13,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":13,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8ea063352ba82a147465","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-db0143672e76f3bc4a36d76a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f38bae38d67d0863d32fba9aef547a3b60a76d5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":36,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":36,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ce0786f440f89630ed3b","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dc1c7de7a6af293fb57ffc52","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"45.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b2ce657fc477b56ef2d125d08d5a29ed238761b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":59,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":59,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-db62357cc90d3ec7d5a0","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d23671606f18f7533ebe56b2","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f9ae7d47c4b59e7e57340c2b682d40bf68b9928","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":818,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":818,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ef8f5061004dc0dad12d","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4929132c181abc9b652cbeaf","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac0e3ca8dcd1d57a64c10666984bcc325060ca09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":795,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":795,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8bc92593ed595cfcc088","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ee005b465cf7f1aa927c8a24","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5772c6217f0daaed5afece493b2d5a2764e5a33","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":82,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":82,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a6bd239c5ddc653f9a42","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f31778468ac56ae44b046e2f","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7db75da60bd7c8fc69e1cd908a2f9cc07c5ccb04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":105,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":105,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-924bd8c98091b9f79c04","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-29b90786a8745b0059745c4e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"35.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b1b9cf4115d73b7bd39249c9a175840f5ce3d8e7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":128,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":128,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b29667e332b8966c6788","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9c9e531caab6f5bcd80b87bd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b1b9cf4115d73b7bd39249c9a175840f5ce3d8e7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":841,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":841,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-486580a033c2c265c74f","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0ebbe5cc20f2c7c29e1c202d","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"be4001e35da3e00f47f857a605d65cb693655b06","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1048,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1048,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd1db4ede9ad11bfffc2","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0335e203a5a0f39b4834a39e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fbadb959934843ba623be84729519dec10b8508","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":151,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":151,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a8a943e3e0131dcd00af","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-33aeb85012b5eaef19d10877","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9166b661dadb01de804cd2991b1cf3cf07aa6ede","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":933,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":933,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-00c642ebc9d1ebf96ad1","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-43f847bc9e7eb4b9cdc426e9","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5f6ed6d09657d5c94fd8b31b4d94f402bfaea40","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":174,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":174,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd3ac34d80652243e9bd","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-880eff553de74a24e587d785","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a2b62bed29f16efb28c9aeed2114f5b2b56f59cb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":197,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":197,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-44f8750b94f0ef37817a","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-52d6339e92b405bd2df1340a","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0de65b54e21b0950b4faa70182e94323c096a53","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":749,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":749,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-03feba1abe578726473f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-04115294c5ecc4387b6f71ca","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"40.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb8c72481dcf0295e82297cf97dd0c1331b24c85","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":772,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":772,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6a20dd5e255d438c1f23","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5733379cac6ba860bf6a7f20","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4661c71a507546c53038f65be5aa1936a26caed3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":979,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":979,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0454c9b0a73455efd9b1","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ce764620b8e6f482a7d0c94c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f1e29d8fe2d7f557046b24aa68f07c25881d8e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":220,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":220,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17bbd23ca85d1af0a184","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1958ce92847a91f4337df67e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f237bb75847ecaf2a6b23c43c454618d9b117dd0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1071,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1071,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2ae519e0c12627071f6f","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6d2f262ccf6126e63ada7d93","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50c4eff32d297529aab9348f749d61f6f8e9a781","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1094,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1094,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3806948532b38776cbd2","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1f0b28b6f118e000a07f99e8","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e3d1e73a64c512a2068f7e00730c62531a6d8caf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":243,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":243,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3c3b5cdbda307482ae34","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-85ae224308cad7fa02613c61","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e4850a044eba165c9c26c57874af3b03dbc22b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":266,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":266,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dc32c945455a401b24dd","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f0f7a8873dfb3a8fed294d45","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb30a71b0d3a58f4cf08ca7a56d542d7b817e7c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":289,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":289,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5c116829788b3826f0e3","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-57d1d01cf6767868e923967a","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f397f20433e16c4b3d92b9b7a856089f08f9eae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":312,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":312,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b994fa880ba13f630c6e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-761b2c58e112025ea5dee028","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ec0be94426d1d2f79f47958b6f3a0cefa4aa08f8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":335,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":335,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ba969b598323bf80ad68","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a5fede18e124abcf938b8ba8","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2247d99f9fd605565fc9151373054b8e54bda495","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":358,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":358,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e0478179b8907fd56597","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f812a0e00d6d6f31859630dd","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"45.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7084d831a1b54fdd04a3e288bd731c5a11e3299","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":634,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":634,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5d683487dee4f98088fb","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-39a6ec4a1aab9527ba8e1583","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b8359f573c6015bff33656b4bc6c9fbeba7cd274","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":588,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":588,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3f015a5fff4d2a98dcd3","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-57e432206fa62b2c0e0145de","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"96109bbb2cb88c0e4fa2e751df636905526d49f5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":611,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":611,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-40ef3fcf9e745a40f547","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0fe35a89c2a4b75b1d8294fb","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9fc659332f62c74d2619ae2c6caa555e1f51fd4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":404,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":404,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-34ce38201cb1ae738c34","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5693f718f0b58e9203a0189a","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e192e5ab007ea9325984115de96e230caef3e0c9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":956,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":956,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9c99dcbc53be8c73526d","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f12509f56f720112a2b813a9","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"55.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9090744d617a78a995daba732a56b1a65e39ea82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":450,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":450,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-230fe590057660917ed1","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4e606045a1abadd543a861c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78746b408c34a377ccb70da27e095faa438b588d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":473,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":473,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4aae4add11ce11f02f7b","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1355f27c185458bdb5f1fb1f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"65.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20b62f4a45b1924130e08ecbbd98ad67184366e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":703,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":703,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-99eb9b6fbad7ef7b9d96","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8cb760649e219d156337002d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"35.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81b9a44540c62d590d8d7d201919d5303cd8ab25","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":496,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":496,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":35.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9bbc3df22fd11cd37f49","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c4255c3d49883b36e15699d6","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"40.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20e5f1e1208d2047de58ac2aa719f6db40551fa8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":519,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":519,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b67bf563ca70954432dd","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7d1c29922209234be31937c7","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23ec7f5f66a28f471c8fc375191e71b81ddae541","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":542,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":542,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-21dfdeaf2a929dfc2a14","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-56481e2e261b269d8827c6c4","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6904c9ace9c8a4702a87957bf3dd50e586652368","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":565,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":565,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ec76597be7504e098dfd","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-447635d31db9fa544e58c591","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a6fa41489b82e39de7bab0eed57761edc779f53","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1002,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1002,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cf1235e7b623b04dd555","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9de4a223fadf9e09b7587b0e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"55c0a8e4ad966f6b0643c4c88c462c5fcb80f807","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1117,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1117,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-python","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aa6693a785d98eedc2a0","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8576831d2a7de67fe938815c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"python"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"60.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c74a14b0f90174b15e262646c0dfb2d4d19fe4a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":864,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=python","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":864,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"python"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ce9e8918188c54f20ab2","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-21f59a9cfc725e551f3a6c9d","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"72.033","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71534a63ee43621ef019fc14cc04a0c82871b881","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":658,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":658,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.033,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a18b8e883dde8ce85889","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-eed2155809411d0cdb135755","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"54.95","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ce1864fb2a18b40b50c8243f779242435d7797f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":14,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":14,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a1c225c5449d0afdb053","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7f68648425373ffa8f226dfd","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"59.15","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04a600750da491ae717ad960163fbe47921baf35","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":37,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":37,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1abfeee9fdb63424522f","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5fc1ce0d3dd432aaca8f0628","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"65.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"773694ee108bb6bebef9da1c833769de47811b62","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":60,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":60,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e9803ecf776f71489763","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-94b9c909a50825b7ae9d871e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"62.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aad2fd7ebd2d751d80a365e30c2f7e6cc3f0d4bb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":819,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":819,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b835b30a74901f39ab28","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5d41e580aa99b5d21a444752","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.583","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a567728ad8d2e4cb2b4783d6a76f8035bc3c3ed7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":796,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":796,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e002c35eb5a681192dc5","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fd24f5709ddce1eb5fddd469","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"62.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"439af34fb8d1cebc7faeb2aa6cfd66cf7f7e54f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":83,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":83,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-93c33935982f93cf90df","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d60785de9fa4748a00871e08","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"60.65","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bff78881156ece390c2d9160eee9c79c58e4bd28","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":106,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":106,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9775f03f0f08a7897188","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-87c82c9f8dd8d3c2ff7e9cee","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"54.05","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f72ec8c859c41eb00169563b3d9e6119ea1a9b2a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":129,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":129,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-831c6a015319b4090916","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1da342a2003d78a8908d69f4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"58.367","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f72ec8c859c41eb00169563b3d9e6119ea1a9b2a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":842,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":842,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d46de25a3c3739b012be","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9cec4cb5ef42cdee6207b4bc","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.067","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"51bbde0ca998ac6418c063f85f749fe38cc54e7b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1049,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1049,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7136e4a26be0928dd14e","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b9fa7d012aa0632574e71ec8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"55.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a959f8b2cce81e4173810c0533a44873f8be0a14","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":152,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":152,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-662b5c6eceea875f9a5f","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5962b0db159f7eee2db3eeb9","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"61.567","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e283ae737cc6c66e1a94f81601d8143af4c1487","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":934,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":934,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0bd9736b52996997948d","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-04eb05f59b28abfbc0f757b2","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"71.05","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f96be0093c20ea6fbc7124ac34f94f5156878db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":175,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":175,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c65927f79bf4092bf731","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-068cef08a7ed5d8b87cae023","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"69.1","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d918ea08d2aab6777c9eafd2d680bf69e2ac8db9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":198,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":198,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b97fd570c71ecf7566a3","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a8496c4278ca99818f697091","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.017","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"167ca0f3f787b012168119a0a08050c02b3c2419","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":750,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":750,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-38cced34b6d83511e643","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7ef045c602af3d0761ddcaba","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.7","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f96b6349f5dd8e2b110e4a782ff0ad34979d88c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":773,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":773,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a7ecaf718d9133f146d4","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-110a7e18bff3d116e23e95ba","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.617","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b1f1053590bac229d6334e065a2fa6cce3614898","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":980,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":980,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-609d23e13b62c6c40dc7","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-899876b5667680c7c9494740","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"55.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4cd30c05f1792f41046b7c5beea92fb4a361726","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":221,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":221,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5f51a000e3d15cfe57c8","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b10788e1c505ab3f2e52048b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"65.85","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"67f82788c3abc04a9530560ad0264679754bc785","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1072,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1072,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c98252d685e304bfc667","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f385df081e0752bd7d8e904d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"46.95","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9005395aadeb0eabbb738362dab594995e865a31","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1095,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1095,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.95,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-82ba85ad4bd040b628fd","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-36543ae82f0099cf64912c08","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"54.083","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0976e14eb9866189648e20f62d98a3204bc6ef57","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":244,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":244,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-24f5ab51938da3e6afca","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4beaed382137b166bed47f7d","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.567","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"433ebc8917551093fc62b3cfc4b14dd45c1dab68","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":267,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":267,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dcf8e005aab115a575de","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-36fa6bb9eed7d7445fa41d8c","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"54.067","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"abd11ebe62f3f98f97eb895251d1c19dd6b6ea09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":290,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":290,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5ee62564625e907ea265","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b61a2ea8ee76c2646cd44333","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"62.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cd7b4ad2a41322ef8cd9128f238c385b3b3ef44f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":313,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":313,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2bae0973733b66d3026b","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ec17cb24dae37413c068b30b","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d210519a612c1af479a96c38e50047e76f0b1acc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":336,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":336,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b2818aa33072bff50921","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7660d29b07e0e8ee5f31bfcf","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"66.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c316ca51bc52fba8f435e2f5bbe763f18fc42acd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":359,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":359,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1aed2c110743b448a0ca","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bfb066a3e8bd972021e9955b","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"58.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"223f6c07b4af42b6dd62a08ef2afe5b94169b530","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":635,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":635,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17e0ce05010b0caddcd2","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-50b6c672e9aa831963f5909b","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.517","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83ed0d23c4ab58088b8ed7b8ac610b8f31028ec6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":589,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":589,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-66884ca8fff69a8d755e","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d7724be7d019afd927793ff5","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.433","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"201bb4304a25046caf1eadebecc2766fc63c29b2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":612,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":612,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3f7f8f3cddedf95484f0","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-99c2deaa9a2b37d282853239","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"70.167","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59ef8fb2c727190bdf6575731b5759a03aba5125","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":405,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":405,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-960fa303a5bfdfd9524e","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0862e84753dc24e24eeaa061","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"69.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b8a23304e32a4da40acb179c570532f41c0fea9b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":957,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":957,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c3db211ff4ccdd38cc63","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f91e7afbdd77d8247bfa3ddc","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.417","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f024a8413c6dcbac014f1f7944389e138159a27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":451,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":451,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fb704ff3cc96d0fbb86f","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8e79635d31817291c8117544","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"49.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba4b493333737d1e9f6c6f029d28a09323208d5b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":474,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":474,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2dba477428529e4aaad3","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8fced8255d853b25d088a10c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"66.367","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7cdf69a7d8e1d67b76ba542953becffbcd63053d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":704,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":704,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9f456dfd8ff9e1b7d0d9","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-612d1abd08bddb8830830594","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.617","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6f16878c0ae13a8058549b45af79058b76445fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":497,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":497,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f741985bdef28a3f4522","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8976ceba9c159fd02635217b","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.733","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6eda2fb49cfa26c9abf8d486107c7d2082b3f1c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":520,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":520,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.733,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1c651ef40f7015300a3c","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-04bc86f50492861560438e2c","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"51.567","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"45d79c3fb22a792531ad4dd17487aef2e14b6849","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":543,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":543,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b182ec364930151a650f","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-985ae2a4f6a2893f2f5a0170","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"64.083","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"592319a8eba9b1060d2fd32a4207d5d8d8503dec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":566,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":566,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-55e4128bf0a7489540d4","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8d164c5c69ae4199c2a1e8f0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"66.7","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d957cf5a3cbf1f532a7a03716b2939e2a1a086c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1003,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1003,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-da5661f07698c5351f6e","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f13f990d461dcc6cb82a984b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"934d5276a4f717a91a74378cac21aec184d4359f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1118,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1118,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-simplify","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7fa5068c84c79f08c5ce","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4237d25d31fe3e3ffe1e6003","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"simplify"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"67.35","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7842d5c23d36467d9fc227aa3d2c84987db4be5f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":865,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=simplify","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":865,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"simplify"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-45b5d75f9ebf5cd37a01","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8fd000a6377dace9307aeaa0","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1ecf4813222e5c5300d96b5f1046d4a77d0c8fb3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":659,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":659,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d9c66715eb068534289f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c3efdcf67e75761134c4292d","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c9327a479027477a871fbcf3f7ae77585775db07","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":15,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":15,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fc48bdd3dff064e7a9d9","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6f72222157d69480bbd3ebef","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d4a2f7e3e783ab1eda8add36671f47bfc241453","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":38,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":38,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-080c7a9b1cfbc0aa7290","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8ae6884b299cef80a8b93e92","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fcfb85e49cc2b01103d4f8a221d6e4f12bb898c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":61,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":61,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-175b4220f0774afb9005","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2f44b92a1390f3f25d09b6af","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9429204d4efb933eb018ff7f0a76f2e4b49d436b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":820,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":820,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b4335936df0a2d1a0a4a","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1ba01e394873afd2cc163e41","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"34978b0e9f655989de97668bd84d9d48e508c2e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":797,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":797,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a87077ca59badfa792d0","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c4474adef838913a4e67b2d3","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5cc2263045a88979024c17f9111b9f33d21faef0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":84,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":84,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-11ab0f76f775302909b5","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-719bc7a368a4364bfac35da7","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c278a764080e9587ab834850621eafbe4901b8b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":107,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":107,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7a4b332e2f65052c6265","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bfadc4f4850a6adaad7482b9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f9d783ac1c48548771ffe10f39099dc329ff3a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":130,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":130,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e78cfc81c703fae14680","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8b87f97a4682215bcccbfe5a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"90.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f9d783ac1c48548771ffe10f39099dc329ff3a0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":843,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":843,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5d6ec75407deaaa71f7c","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72f264f21fe16c6f30b15e31","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd584c80b639f065ac0a1a034ec32c71e3c2ebf9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1050,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1050,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-661f0ddbbf63c61863e7","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8f0cf2c414211cab0d9f0575","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8cafee8502e4b7ba7fecf7ae4c7dde6b293f4925","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":153,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":153,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dbe640e9a1b528ca32d5","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3e50a6b7823ca87db1a42cdd","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de838d9c49db59244ebbb68718fd434c894c4ce3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":935,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":935,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f2dbe39793f5f3795bd7","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b183e943cbd66d26aa990b4a","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f3abdec30b4036924b0609851debe3c21ccd2a43","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":176,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":176,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9f57ace68879c46f956b","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0ae3e1fdb5f9e21ce37465b7","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53dd0d5d2ea3b954a8274844b8d5fba047e27893","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":199,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":199,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-64f16d4b26c9b7a61efb","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1e3ddf91d56c57a0470bc4ef","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e301923b8b9d02a57e0a6eaa7cbbd87051a6cf27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":751,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":751,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8ae723a1eed6d3584cdd","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7e8fd2fc85d6b7efb60c81f9","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58fcd0ff0ad42969fb753113baebea99c980c3c2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":774,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":774,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5854b948f313d4bc340e","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ff3787d417f50d06bdfa013f","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d7bd62caaf10ee55834f8fe716b54464521592a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":981,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":981,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-72ae535c50b8cdb897f8","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d5706c432f159e46ec65329e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54ce193bef6456608b5da11ef87ca569ec8a6384","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":222,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":222,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5f35f8142267c9f8a96c","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-563ecf217afb3e790c597459","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ce4d2f6c12aa613830677417fca766c95abaae22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1073,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1073,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b1f56d1b191910eae0d2","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a55fa514246a27999de4d7f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f45b4a5aa688f33a6461f3f0ee5924041eb46fe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1096,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1096,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1a42bc5e507da972fcb4","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fc1b8a7dd7fc94ec82b4953e","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4940706917810ed98ecdc51b45d2b48dac187db5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":245,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":245,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-60eb57a083def55a0923","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9928ec256ad633d9643941bf","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ee29011a912fd9b8b2c1f93299333fdec748d4bb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":268,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":268,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eddcb8da630931c4137c","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cb1e0c7c55b20810e3f4e45f","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94a18e2a789e023f83190c0e3d41cdbcd7d46232","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":291,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":291,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd81a4e3b550d3dede76","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-47b3afea9d180ed6596ae6c1","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32e6093aa5394f9183d8e00e829f50c1a465ec61","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":314,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":314,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9c878f5b0443468b6cd0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1abc9d697bfc5d525237b6ce","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dc14a1509917b9cc7038a1a79f23b1c934355e27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":337,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":337,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-269126bdfd360bd4207c","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a8951823249783f75e29d0d4","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4a7abc54ae6d08292a5bff47c8fe147eb4ee84da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":360,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":360,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-12c3f5a4b5262a9778eb","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-af890d32500d01714648b0cf","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"18a12c6fdc8e22425ea733d40cd2bb7b4ee632e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":636,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":636,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c255fb64191a516c58ea","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a6712c764f977808e8bb7f7","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6aa2dfabf991a2617fa81203a8d0aa29c32e5c21","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":590,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":590,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9551b33f430bbaf97c33","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-775efe9497b09ef914f119c7","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04692c90ab753856419c6836ac1b691fdeb49e22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":613,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":613,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b0ab5c47900059bcb96a","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a3a2a903694db9a4cd578d59","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3bf455451053b235ac214a6ff960030da135bf09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":406,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":406,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5c585fb02e0b30b042c2","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5b7fe67cfcf9e35176d155e9","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ee7c4f1eaac0031c1863f37df52929a9e51d371c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":958,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":958,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-994ac25739c0d0ed9c41","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c5ab844070c634686aa05199","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fd140d05e26cf93c395177578ab6c6a492febbb5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":452,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":452,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c427f48297db2296fd69","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-24494b3252ad76af20970f70","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"92.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a5353d138d06a4b96763c98edccc5385a95872b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":475,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":475,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8ffc90fe57f5333ce99e","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7a86a68255f9200e419a3d74","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60c83821ef38198479b307b35088a6ceb78e15cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":705,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":705,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-21761fe8c1df5e908260","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-175907b4e1db8c9065aaf750","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"360577b4738c6a5053dbd67615b1c9b24e7394d5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":498,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":498,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4c18a695dfdc2613979b","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6023b073f8551ce5285c52d1","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dff679ac78fd1f580c026c47aaddd9ac02fa8ad5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":521,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":521,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d38c180fc0af8b6af350","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-89bfe706e82d1e04960219ee","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bc36075a4c2aee1ff4d12bc78bc2d7ab3035899","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":544,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":544,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-01846f809daadca1c153","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ae036f5cbfcde33d6c2c05f7","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aaeb599c0cf480bc1a2e781d81beaab5a64db603","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":567,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":567,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bc75970cc536c7317bc3","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c88b855e0de8b91ea03d5dfb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b40d01140acac0fe497a42b1183b4b7c9ee18837","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1004,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1004,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3899747186fd16f9fece","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a8ac15535f7b6cf6312ad2c8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52695aa6ed8fc760249dacc87525fa9255b47eff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1119,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1119,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-spatial","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2baf41b4d1cf7d1afadf","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ca9c77e6a059e424aac86741","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"spatial"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ece0c9ad1b19b54bdec5b21b91cb50d7a29ceb42","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":866,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=spatial","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":866,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"spatial"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3f1bd469fc38d7b489ed","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-45591e44f49a50353faad1aa","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"71.7","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5194d1fe5b5fc0a0909d471998f026a23632327","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":660,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":660,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6fd73c31bce44fa9eb7f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-702c143843acd91da470fa89","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.85","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"33ac1ba922e5fbba69f54bf8df47057e0f1bb147","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":16,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":16,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-62f58a9d5352e7d19971","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3af2829854f825d3b7c0845a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"66.133","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"084eacdd9036400aa84fe57914d41b4ddf88fbf6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":39,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":39,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-db92563e80546ad36633","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a1c336b7ae254ac07b649b2d","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"69.4","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8774df188c1f952f95509f0e3632a33557d84196","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":62,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":62,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-31fc14c763d9811ba75c","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7c79cdb8f804d4d4acd7cf35","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.6","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"da506eac8748813159d51addae3278334bcd74f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":821,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":821,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e1401426aa0ae6d59192","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3522b2428a0243eaf13c5609","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f69ee9ed7cdbd2f2347df01e9df8787e3e67f225","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":798,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":798,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-23c2787de3aa0be57e05","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-eab9164bd60c9ae20669f8c1","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.9","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a61936e05fb3aa42dc80fe6931f4a91b60b5b12","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":85,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":85,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-835511817c4257a23482","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cc1dd3cf112a504a3724fb5d","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"62.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04cd17886d03995684e559e1397ae0710bfa3aa4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":108,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":108,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3c469d18ecee588deec5","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-56b446bc152be6b429a53a11","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"67.267","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92c51d5bb2d7114c7e5794839d88f349f4e3e1b3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":131,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":131,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bdaf9b274042f180396e","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b35e868755fd55cb7267c08f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"70.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92c51d5bb2d7114c7e5794839d88f349f4e3e1b3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":844,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":844,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-933dbcbfb7b6ca28fbfe","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-461aa23fc10139b4a4b424e2","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"73.833","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40a64e45c6ba42afa338c08d8573d5788ab2d747","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1051,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1051,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d911c848c240ee063032","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-96dd6829d9f59293c33457a6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"67.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f4896d43ec343effe219d6f6ccbbb525d269bca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":154,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":154,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b14414bb3350e4567aa9","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2bbad1b16bdc23d8af4e3d18","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"98a2c85d0c4807d96b88079748abc3742edc2d15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":936,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":936,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-14d6aae49f1844e3d133","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-499f59eff34653e02564b897","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"77.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3a0285e92168766cebe528a80b040342d1296a9a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":177,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":177,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-97d4d13a1db6792a6f5d","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ec934995d4ae7fe1e062f0ed","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.2","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a60e75cec5a6793e8969280c60b36ff99ab186c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":200,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":200,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-550c1399b28297d23fa1","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7a427ce5f2f98afe8a13a050","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.467","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9a5b57e90405966612bf33e3382759dcd2f0dcd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":752,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":752,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.467,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9de56105d9dfb7f05c6f","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-279861201a52b3fee00f912b","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.6","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6341ed8d748add1de64208ccbb24a35b9b484fca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":775,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":775,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-444ed62572ea13420020","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-61385429b88715d2dc57cd9c","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c047495342813e40802602c3999b8b23b154c7b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":982,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":982,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c6cc5522ca9add749743","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a828791d269123854d9d0df","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"62.883","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2183e2a837db353dd5e9193c02578662bfd4dd8d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":223,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":223,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.883,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-63a7997eb932b74b236b","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-21fdd711dbe7c2d2895807e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"72.133","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"01c76f6204725c888a2acd705a3bbfa28b2ee5bd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1074,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1074,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd6a80901e29dae3b594","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-417cc60766ee8b6401972723","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"59.883","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23537fc0971a7c4691ca6c8c6b6f9733c3aa9f99","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1097,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1097,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.883,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bf0aebd1e210ff302af1","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-10ee948443019a84e38b8582","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.7","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e508a73be383c5e37128279565f5f54a76888fc0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":246,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":246,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-095063923fc28af12b89","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-63ad00aca489b0c8f8fa42d4","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.783","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4cf0bf200f47f10fb57f34f2e00cebe1c8d1929","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":269,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":269,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.783,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-00efae3087f8e6226820","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ef814f8cc9a0f97924968067","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"66.633","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1b594cdbf8da92204d21a4ba1a8b3faa224d4e33","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":292,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":292,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.633,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-64f040ba9d8cda9d2ee3","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-27c981d9e0e7e680e45fa9ce","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.653","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b78a994ff7fbf4297ddb8391b8d18e85ef2e4000","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":315,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":315,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.653,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-53ecdcddf21ade0419fe","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bf6e07e92e918fc40fef0156","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"79.017","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29cfdb007e035e1e82eae8c947b9f53a4f338fec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":338,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":338,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.017,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9e71b6cc328447c3a4ed","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-559ec8bbcdb03b8c2393c7cc","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"72.55","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"855aa927b4bf7c1847631ffa721ed2e5e9720de7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":361,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":361,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5dee263bf55a6765e1a1","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-afdd7b4aec16a5b3a94b6d13","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"69.067","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53bc26514eaa866b6a89034f5d327cae9a6e38a0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":637,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":637,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.067,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2dd0f9b9367678cf4a02","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-045a54cb7c2dba2a8b25e0b3","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"75.917","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"af185b0f8b3d8c0542365713437d113d971c07e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":591,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":591,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.917,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d2ceab4158ed2d234166","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e6131b2bf321adc400b0c454","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"66.833","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"00f88693b04e40a658f8bb7ed26d33bf34c6fa52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":614,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":614,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.833,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-47d9fb3b8f79985ecdd0","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c247c44464d7f85224bbe4db","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"73.583","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ca2a78d016cf3df19c3cf8453bc30024426410fe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":407,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":407,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a236c6ac3307fd5170a4","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-94c0389c9a409fac7978cfd0","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.267","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9c742892f31d3628a90441abc56df305004681d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":959,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":959,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.267,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4732b95a643652d486be","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-43bbfe103c07b9ee809aee3a","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"66.883","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5b0880e69ae48435f6bbff6ca71dd47ad2761721","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":453,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":453,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.883,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ec4ed9cb97f54dc7d3f6","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-207bd018a3585a03c05a4a5f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"68.567","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04eb81cc0f53ccfdcf7515001028d6ee7b67ac97","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":476,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":476,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.567,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3bc28697d1b85c23602d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-144faceed55b74ae33c5957b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.367","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adda85bef075a5c9ac23ec2cc8ae5d72dae96b77","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":706,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":706,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8c911107adcf60fd2923","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-94dd8ba4b34b5cebb388c7ed","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"63.55","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4bfe1505bdd9b916f93abd16ca845de21a2a145","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":499,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":499,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1787e5a069369deeb7a5","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3f2630060dd6a78c3c004116","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"60.133","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37ee240fe9492075d6cde99bd0a598c63ae7ca14","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":522,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":522,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.133,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-12a2d9f24e5895e1418e","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-691552cd94d5d8a3caf3a17d","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"56.867","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3d317eebaa0a828c83c111cdb8c978a5efaa40f6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":545,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":545,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.867,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2698358a6b55707d10b1","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5c88bab6ec189c3d9e4c4121","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"76.283","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a4e51d287ea8aaa9fb21c700f391d7b1972f29e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":568,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":568,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.283,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-492a2f1ce08c5c63fd92","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8cf6a05aa6e0a39a6a513254","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.1","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"78a14e8da31bf1bce31442f865653b582ccbc2cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1005,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1005,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d2fc2bd58283f32aa9c3","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8ab7f86b3780bd4b53678f86","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"81.367","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"877651b61c377713b007107c923ee63a0f83c28f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1120,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1120,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-story-generation","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e13b0192e1345ed14ab7","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b25f3ce30c7033d3efc6648a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"story_generation"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.517","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1b4ba624f1ebae50963b2d8ba32b588ecc0f451","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":867,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=story_generation","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":867,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"story_generation"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.517,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3912647a67a64abc2782","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e8e5ca2a675775ed9035f5a2","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"81.7","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d58a10788edcfaee09739bc725f4c8a5191b8ace","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":661,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":661,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":81.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-90af9ec2797c147f1483","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6d1ec17d5d858c0b33b01c41","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.717","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"215db23ed34bf3be2c47879fde7e2aaf08e18c0e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":17,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":17,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.717,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4bd35479eeae2374cd5e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4ced612d9ec3bbcb8b147797","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.15","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9862796fe0e78f14db202bac8fd50c16492e7cc7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":40,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":40,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7ecab4da35747a37a709","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2e8461aae355f0c7992081ae","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"66.283","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"407f0d10e2f3658a8d96d47627362ebfe44930e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":63,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":63,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.283,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-607eb0dda6f0c025138a","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-607ffd34242c657c71a7e229","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"77.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bf28f130b3b06f5ecb753f50f705a5f09747bec8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":822,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":822,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f5ec376ed26b627761ec","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f1ec528c7721646bd318a322","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"66.433","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d66dd0380e94df430f8aa525962fa33a538a940","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":799,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":799,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17a4f3e9eb7139d8a70e","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5f4bec9673a497e2b8c42369","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"61.65","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0719c48b23185114f8669916b6be78b7130954ee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":86,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":86,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.65,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6e8e684c0babd4beee1f","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dc2285e84269cbc3bcc176c5","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"71.217","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8fb9702dd3dbbeb2749cce457eee2d324e7be98f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":109,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":109,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.217,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b50751b354c607330cfe","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a4501e8a38a7eecfc59d9258","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"68.85","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04f133beb2ea48a082a66e1ef717dbdd1f7e9966","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":132,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":132,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-138af268c6188954bd78","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-858ad5b323b8994c46e59eaf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"71.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"04f133beb2ea48a082a66e1ef717dbdd1f7e9966","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":845,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":845,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-690d1126e954b2c6007a","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6dced55154cd239608ca41e9","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"72.1","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"754603a46046f6628a6e074b6f3731e1c66f20a1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1052,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1052,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-70e82da89076617cf389","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-abdb37f5515ac616ef6e4d92","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5b2ae13daee6f6bc96c441e4ab329ae40be4f24","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":155,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":155,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f1ee1e98f17594baee44","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-786178516d61f4981a92fbd8","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9653e873466202b72237e9a52672a1e63fb0fc53","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":937,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":937,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0e7167afa563999a0535","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-302bbfdf6ed7fec358730e31","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.233","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd749926271a293ad3b8886e678586bb39d1b81b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":178,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":178,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.233,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ec14978835b0f2d74435","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1d0f7606ae79177455e23a77","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.583","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc149b143f85bf5ae8e2d745351fa95ff327d11c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":201,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":201,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.583,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-22076d485174ac230651","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8c0476f1a69e592776e4b42f","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.467","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4db5dd4689575a427410947448d2dbf82c462418","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":753,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":753,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.467,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0cf408e3f1d992ef61dc","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1dfe5e20d4f89e9522e62d76","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"75.85","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"760ca6da2ef9dc7de08636fbda5524fa61f40217","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":776,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":776,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.85,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e8ff260d5d8ed3cfcca9","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-825eab9e5721966c812af04b","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.767","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"660424b1ceb07aa4f064bb8769deee735708b54d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":983,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":983,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-388224aff4cb38f0f756","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-aa63f60ab418f3f8fa0e9e32","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"68.367","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f33d2e685d50ed82da90ad9c731b6a1797cb3326","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":224,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":224,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.367,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fca6d71e70f3add1db4f","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5bf09074409a6965226b3fe1","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.55","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"01b9ba7af57353b70722328cf7d5a43cbadb57c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1075,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1075,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.55,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4e890536b3d0e736f22d","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-580af7e188d2ab6979396655","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.467","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ada3fc6a6be3915dc35bc6554a7738e05ad49de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1098,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1098,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.467,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6586a82e1970764407ec","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7ea92aea2cc9f98bc4ea1efa","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.2","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6fee043a9c3a6a41a6b22ef8c5126ef0c491005","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":247,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":247,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-94fc38d9bed4ee4efb93","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7622d4e73c47b297d60d578d","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"63.383","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2a2252e25ba6c72e0f55eca91ec7c6afd50d633","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":270,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":270,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.383,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0eac23b42b3695b0735e","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b526081e0c25bd7e76109e4e","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"59.633","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a87763a62fe46472e17c966d3e16bd6ac1eea751","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":293,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":293,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.633,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cdbd6f2949d10484ba89","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6c28880daf6b19b5494e0cf8","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"71.767","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92f61f5ade1ad49857ea169fa809b494b1b4a14f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":316,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":316,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":71.767,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-652e2cde5139b1da85c5","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b135c4f6cf429e8450dc1751","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"68.417","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1974d0a833828260bfb7a2a60246ba16377766e4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":339,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":339,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.417,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-06c66a22acc6cb4e7aef","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a738a99a22d5cdb1ccb26ca7","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"72.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"690b1953e78c733d8efa9b7f3b97be8fc2ce7219","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":362,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":362,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6b5d59233e5ef733483f","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5f78ef5babc08023f1514f5a","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"64.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d41fff5b5050ef123baa1b42c2ddd41327c87521","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":638,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":638,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2c3c1a62100317b61c84","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3b03360e41de70b5d6f3a1f1","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"69.433","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e1cd95b1731678b973845ebfe8b4e4159e834d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":592,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":592,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.433,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7a6e43d17c5b09b422e5","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5c5cd36e180200f5c4a6d9a9","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"73.167","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f53512925864493a56db8093c2363e9804633b46","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":615,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":615,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.167,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3689ac12e589519c7053","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-905aa659c10e4a0e96d32587","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"73.15","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c1d9d02445bb3d5e2c11c6c98cb63d9d4efe1808","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":408,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":408,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-150fa58cf82478041f4c","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1190b9252bd7d319f290a0ba","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.083","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0ad1efc4ea30bb797595be58d7a2252f56102a7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":960,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":960,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.083,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b051503478614f927e29","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a5bd101426a746d8f5db197e","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"67.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0cc08055ca2f5aa60ac10b963fbf6572d1b5aace","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":454,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":454,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3b86e17572bc6165f839","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-21daab08a8310a71bbf4c50c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"58.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"28836f8f80efcc92b80e40c08a7e3918bf56a83d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":477,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":477,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd1a88ea0090d59d7f09","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-edbf2d5d45fd15619ec6ff61","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.35","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e61ae0e0f2fa0407163a38afb280e64e885cd34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":707,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":707,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3e2cdf3bcefe7c2e211f","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fba6858ab982e0c8a8084f4c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"53.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b004b7635387db1836bc554ba4c2572e4cc4b9b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":500,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":500,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":53.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0269473cdee302720094","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7dbb7d2fdda54e64fd746528","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"52.617","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a6a09a104d93952027538cd9b7ce7d77e439242","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":523,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":523,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.617,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8bb65cdc56d909b4fb5b","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-857f6468ad73732dcb8d8aa3","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"64.317","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a63a187ba6b26bb5bae62032d623f40f55706bea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":546,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":546,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":64.317,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f6edf3af7c1f01e33d2a","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8bb96623da5c0b5382465be0","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"83.8","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc4bed9d2f534848496baccef8d1f2181c6ef504","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":569,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":569,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":83.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-135df57c3cf52d3bf8af","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-567e2c8a7397327d2c098650","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.933","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"461b17f56bda4bc548d61b06322a5ff98cc909aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1006,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1006,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.933,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e2af57be1e883f12b2e6","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e3a76a7a07d0001c4010c4b4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"77.35","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fcc6585d5ef7c81eaab0f405376ec355a514a1f2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1121,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1121,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.35,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-summarize","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7bbe8409329f7bd4f76a","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-35f5b20384ef548b37a12cec","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"summarize"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"79.15","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f4ea22d6c3c0ec7e81767c59cf7477bf7396afb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":868,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=summarize","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":868,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"summarize"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.15,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b9feee980dd3f7234fc5","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3dcc80b7689e250ec79fccc0","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"56.135","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f61d1f8da6530fb84b38dfd07e71a073f435db4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":662,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":662,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.135,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-665d528b024fd55ddcae","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1481dbfbb1dec8e58c82875a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.923","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"11b6b17d7cf48fa38ae84c224d2219efb38dcbe7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":18,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":18,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.923,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cedc6ffb20908bb614e5","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-36f9c9835edec73759ecfd2a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"48.173","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f332f3a9e1ba3157877c820c1e3912a487f0e67","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":41,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":41,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.173,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8d730278f071d207531c","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1e02ab8350e47f8a53f8f550","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"47.231","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77bcf6af34ef2baf074233bd1781eabc474ced88","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":64,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":64,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.231,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a5236cd73f2536ba055d","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-17487fcdd86fbe2506ef3e8f","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"51.077","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c5ed7f4e618aad60d4498d0113acce3325b6320","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":823,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":823,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-89d89d8e082d1fbc22e9","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7e88576fe8ced687ea87af84","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"51.962","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b7dd03ee7fcc51e9c5e6c9367b028b282d3cb1cc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":800,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":800,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.962,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-67baa5f6f94bab81f4a6","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9db6b5caf4324b9374cd22e6","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"45.077","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eecee9b9105e6c4f180c8fa1aaca88f88065260f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":87,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":87,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1f4caf785a1054cedcaa","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-045690153ce3f2e7b41c0bff","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"42.442","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"12f124826aa49344ebb0256036029e3392d05314","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":110,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":110,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":42.442,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a451244b6262ff91e200","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-27f4874fcb5cf6ec562712e8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"46.115","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ae48c69c962f0efa05ee434bea7f91b3c26dedb","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":133,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":133,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.115,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6b85f74b122ebd9c9288","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a85c4f6f876bc5fadd47d7fc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.538","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ae48c69c962f0efa05ee434bea7f91b3c26dedb","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":846,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":846,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-26fbab3ea415b64a3ed9","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4cdb51f4f289e6fb4c38b7f3","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.346","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fbfdfd9fd18cfd4c2abec81d8e8cf56c50cf891a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1053,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1053,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-32fb5c8c2212131a8e93","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f5e418470acf15ed1bf9b0a4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.192","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b767b6ee36baa6be78732f4fe315c7cdf695be11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":156,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":156,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.192,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2c06832052aa2af5efd5","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-695ed9feb0219aa3f7be9404","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"49.462","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a08fcb47f78fc4c243bb7f49b9a361ff758186ef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":938,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":938,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.462,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-486a255b09bb2bf1ec07","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-80938f578993741e90d4b0ec","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"52.346","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"627faf0020f0bf23203436d0a59ad653b614a6a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":179,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":179,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d4848ca09488541a865c","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f788e67f1fca02d3292ee8cc","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"30b25eb4fc336dcbd8c106f9dd0414032a59980d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":202,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":202,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd097e88b901833908b7","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-02651e5fd834c9cb619d7aa9","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"40.615","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"77d6cb7ad515a6b5f0ba3e3525602ad60e749d31","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":754,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":754,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":40.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d96c07748fd9789ab045","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-17dee8829c1d8df6ed97c629","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"47.135","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a5d859da4aead4316cfe6828901d0b026556b302","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":777,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":777,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.135,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8158e4b177b4512c3885","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8787ac97c5108a46c5d4d616","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.481","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"56cc0b9af96813347449bd8076752a3db1a09627","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":984,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":984,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.481,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-021c60a2085b73366c67","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b8ccf63f6a8a80bb963d2e75","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"45.865","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fae1bda23c6d2195606f800ec76f9945a0cd1b2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":225,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":225,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.865,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-46d2f0437023c7c700d3","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9ff9ad8dd2350bd8ff676831","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"42.346","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e0b5d7b8ef1dc2e3d5a61763d387602348b385bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1076,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1076,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":42.346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d879939bb1d192e10b6e","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-71903b464be0c577c46b6075","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.827","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f5659cbef494216f68d31b957f0726d8e60d235","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1099,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1099,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.827,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b5b09d6ee32736bc67ee","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-69d1ec10ae03519fe271c548","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.577","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c7b060dcac1bc229386cc9393d2fd7b95d3430e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":248,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":248,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.577,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cb8e0fc22d5fb85fea47","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-80f2652b6208d6203539fab0","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.596","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6bff8f885f4606291f597977c6038cbb876b9813","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":271,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":271,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.596,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-40ce22441187ecb2874b","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4289c1ec6a3486e205d93ed0","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"49.385","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"604833ab9b967434ec0219eb248c19734985bdee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":294,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":294,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3f12e31ed3043b1b1312","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-aab6f1b077c360dd077029c7","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.588","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a22a64c29bfb0816024f89a733e0fb1ffa1f9c36","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":317,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":317,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.588,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6b4c15131982ddedd2e0","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f100bcf8e1d8b7d9ec71379a","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"51.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ae2a2f3e8b4f96da40248c4a577d4bfb2f0c1070","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":340,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":340,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5870b37e4c73c492927f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ce4022eae9ed5c0e8ec3630a","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"55.904","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c4f944a0ef9ef18b213b5db3b450c79665dd3dd5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":363,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":363,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.904,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6183694589f4ba361860","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ab60e300f64b5f3a08d87a9d","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"47.596","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"61c1a1928f153da6b1b579e17c5b8c1bd9a657ba","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":639,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":639,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.596,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-35a3f35f5962ff5e88f4","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dd5221929a54b2bec155dd4e","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"49.308","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b0c96da284581b8ce42ff7fe27d5484586e2aa29","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":593,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":593,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.308,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6a2e3eb696a0d79cfe30","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c57a7ed49b4c0be5e94ba413","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1f6ae22d28ad45f7f8a76c5f9a327d386f78f851","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":616,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":616,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0038e1cb8f38d6a29447","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3d53d85d996ac88391ea2f64","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"43.577","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6b4d8730c5b7f5de3dc49bc3ea22100703026ebe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":409,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":409,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.577,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5a8e143bab105ff4d5bb","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bd5f028a06827ca1e7fcfa94","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"43.423","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a6c5966bc97b34f933064fc6f3f736fe2458ac3e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":961,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":961,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.423,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2cd998c506b99c817e3e","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-432757202a6e46acb53a21ad","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"46.096","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c290104eaf7e026090ebc5592e63e372c6fa0bc5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":455,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":455,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.096,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e96493b7454facdce593","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7d5f525bb575f2cbd797bae9","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"47.538","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a35f4bcafeef951d4c05c4c03f6c1ca25b79d1ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":478,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":478,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":47.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-03971de9c1a9a897266d","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-15cbcf4ad4313775499bc1c6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.404","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d158345248820e9d2f65c8029128b81481a58a93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":708,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":708,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.404,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-10c10d4c8de7035b341a","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-54071ed178f4314df860705e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"45.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"42b3eba745346304843f00472b5764ddd29ca628","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":501,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":501,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1edbea0dfea44f2d7f09","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bf57a379c07d235881b65b73","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"42.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f52dd29f6c3291f34232ba8c184633680bb1e229","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":524,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":524,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":42.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-87dbccc0c3e525827be2","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bfc53457b8a58ab19df64a23","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"44.654","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5e7d5c42bb64d0c2bf544801e854fd7839dd0ee9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":547,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":547,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":44.654,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2d6fc9f17f597a4337ac","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-43e9e74a9a0437a0896c55c3","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"45.519","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"03b5d3fac87f130d71fd56a56b85dd9de41adab6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":570,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":570,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":45.519,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e9253c1d946be984d94a","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-92cc580865c38d195fc3fa7d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"48.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f0dea4e754de8b4ffe487f9f30655c30d5b583b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1007,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1007,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":48.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-482401687db3c6fdb74d","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-030725f4aee61ad3520aba10","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"51.019","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6c7dfc9a195b1a44eaa8a6642924fadbea6daf5b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1122,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1122,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.019,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablejoin","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a79d41ee00126e4ac2a7","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1e0e16be541530b68545a86d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablejoin"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.058","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ebaa32360c6081a4914ccb487df04e29c35804f0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":869,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=tablejoin","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":869,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablejoin"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.058,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dc5fe36796d6ba4918ba","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5d686df9a5a9e828b622d2a6","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bbff4bd5d70300b6aea75b7305272aeb1642edcf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":663,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":663,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f9f497b00a7435005618","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-91f97ae94f2e63ccd7789dd1","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93c33594dcdeb0e7b331420c87a91b33ba25a330","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":19,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":19,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9d63200143070aab57e0","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b9c5b72858f1745ca1593deb","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25c33c48ccec1b1ffd30871f9b6acd260cc0a233","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":42,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":42,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0ada2dfc5420366616c9","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9747a7d33757016e0d508bfd","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82766ad2d62c3e6dfe7bebeaced42d7c86862fb7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":65,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":65,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4d5404be5cceccda5099","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-917bcfd18751b92c775ae1af","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb1880a7e7662dbc35f16ad7b6b1e50d28cdc466","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":824,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":824,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d1f2dd36a2473e858924","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-405c28d811b05874cb873e39","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.118","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b0a31e569f510261314586c86bc1b368c9f201b4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":801,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":801,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.118,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-987e724dbab2deef5724","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-373b1f5c44253f92bcbc0887","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"05001dae682b7726982f0e00337e6cfb37d0816c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":88,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":88,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-df0baac66e4b242cf01b","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-180baa7ba22a07f652b26255","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b59f742f5e33d3655506df3afafab0745150a9e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":111,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":111,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4d83d822dc1fe09f5d03","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-539521f6a9229abed2475afe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c77752b26aa7eda6c33b30409ce7eb13d1d0156","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":134,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":134,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a878e7f83b08e5e5f7cf","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2250e2f4309199e417530f19","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c77752b26aa7eda6c33b30409ce7eb13d1d0156","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":847,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":847,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6db8c54fa02c0615d2d2","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-48a8eb4efd54fa4f29039fbd","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4132121ead01bef9e8fc35cf3f29a440bdd49a6e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1054,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1054,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-08f6fc46d6ffdbf3f7a6","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-647334cc7eb4223b879bf9a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"546e1aa81a29ce5610c21d90cf654c273bd336a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":157,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":157,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a28a7141b32f64da31d6","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3e4d59278a5df7d1b3b66e34","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c501ecf0ade5062921bb4f1345c5387a53e0ff0b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":939,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":939,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d271471f5c0a39a7e491","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8c371173111cd251dbff742a","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b958c6af419af2b833a36667aec0fee7b15d501","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":180,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":180,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ec4cd1fbe59022e6c402","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2c92cf765d6e8b161758c553","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ba639fd6a7e70abda7ed660ba43bf46cf1dc848f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":203,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":203,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-675773553d972d950ea7","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dd34def05d8cbd69cc0da29f","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"863bb99558959b7c8d02ebb0c2281200283322d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":755,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":755,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d72c1bfbce1e02e80d13","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72b75648358753856ce5fa9e","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7b316968f5ee91f67104538183a4fc1bc6b1f186","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":778,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":778,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8e81ca2d8fc5e7ad687f","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6c0aaf46b385029e530f38ba","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e186b452cc329dc0340bf9f24f62df573a0e560","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":985,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":985,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2be29c9f5cf657828b14","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-88495e452dabecbd2c1ea083","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3f6225708f3319dd60bdee3627067f9ab71db75b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":226,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":226,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f0fc33ec9740073b4cc6","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-68320be2c4463d8ff862bd38","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"70061129525818a365b55d93a45898c8966030f9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1077,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1077,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-88e94e8166c55fc420fb","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f6350824bac8e9f2a68e2843","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb20c5dfd079c84d0c914313341ed188a78d039f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1100,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1100,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fcc77237035d4e68092a","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a802a3be0bec145ce169506d","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73db1fee647a0ebae5c78b512ef064c9dc45b57c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":249,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":249,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8733f4423c5e69f7ceaa","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f70982ed1b0c05a15fee1a67","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"015e7924ca9e837925eea59c347cb749f774ca94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":272,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":272,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8c1323633eb6b065b80e","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f924be6e15ca66b36360d371","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"10b6a90570559559f1f60c601d6cd135cad67d21","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":295,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":295,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eccd4abde519b8da6046","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-70e9d99d3acd42f6ba8b2f8a","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e21243ec02c4298e23ab7db51a335744a215e89","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":318,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":318,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a9a6c3ebc4d1dcb087c9","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5b0870b059206f45fa417848","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c0ff4fb67caa463ea835f9bef2e9999b6904763d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":341,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":341,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b0333b6a4ec15cc7be0f","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-708706e7811da6408cc85f79","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d53218a1f290ae42b3b9019bd2d6c87f5bceba6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":364,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":364,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1360ff19968d79cd36e3","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-144c036c92e8e85a513fc348","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4e8ce8d54b20b4f077d1c8e3cc094b8dd56f5871","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":640,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":640,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-55c467bd493992a205a6","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a7cb4a7360e67b5798434a20","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d959d1c0d6e7648cb11d6b3403e7365000b77ff7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":594,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":594,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2d0c818ef39e676bf934","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9f2dcbef41765b864c6e107c","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"29c97ccd711f8119e16e10e3d14b3f90c4af449f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":617,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":617,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1b5ce1d2cbcf5356bc08","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8d0b0ff165fa6a13d7ecb5db","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"33e4e559d62b1416bb4cedc3ee7565f11c8a28de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":410,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":410,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5ccc84ad985c8bdd138f","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-03fffd31dff49eefc9edd86e","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2275dd707b84c5a288b3afb272366daf6cd7cf89","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":962,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":962,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d1fdb65bf123854dddf9","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f8eca8a3ba8156504a5a97f4","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d31ece7d8680d83cc812874bdb90f5e067c03bef","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":456,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":456,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d2db0369e3510519f39a","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-796ddacf855f66be7c523bd6","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"92.157","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa11658374739d845cf7dda6b94be74df51cfb8a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":479,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":479,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.157,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fb5ea82bbde5e0bfa917","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f1ef46d018f172c60f752c1c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5885f4e90041efb743a146a0af5aad7ef54f4a9b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":709,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":709,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-bd669793d4e5c77fbd26","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-07fe3c5e228af828f38f68fc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b82138f868657224715f1732bab15fd16abba226","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":502,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":502,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c9a44b71f93aa601a2dd","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3ee416bcc2b78119d0207649","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2ee9727a35583893de0342926573e07485f856f8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":525,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":525,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3ffde5e92dac6c8cc4a5","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d85cda69cd3d559e977d6c76","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f036bcc1f0eb0c8d0a160eb2cc7186c9129f6e37","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":548,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":548,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-95f1afc08dea872345e6","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-01808d297c74a60d57790bb5","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"39f6f1101ccad63c54a53e7c51e5107441ac336e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":571,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":571,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-193e7712891d4d39d709","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-27bf4b8029be9f98876084bd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"425679489e5c9fb52b91c32554b2fa0f72a289e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1008,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1008,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-140df455ee112c1aaeb1","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8ffe1a0f6d863e65ab5925cb","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.078","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"67dc0eb1f111b6ce782e6365f63675de18e38635","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1123,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1123,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.078,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-tablereformat","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-828b0ddd82df6d597829","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-79b95672e255debbda32cd58","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"tablereformat"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.039","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d96699c7547bb752150572d38085c4aa36ab151c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":870,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=tablereformat","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":870,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"tablereformat"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.039,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ef964457009a6a80169e","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-12f49ce0f8821b18ee3257c6","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.615","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"415eb415b9e248be9b399179ab2586e6cd4b57d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":664,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":664,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3be55b6ff5aeb2a85214","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-28bebbc687c0894248a134f1","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a2e7039fb41afe9f91e265f420b1799b86557ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":20,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":20,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5b9ef22b4e7ef8b12b7a","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cdb7f24f9c54666dd295a36b","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.692","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d0d133c2c63d6189abbfe7da1ff4a8014d70d361","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":43,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":43,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-db2a6537de5bdcb95f5e","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ee6149d0802449eb2ede2e5b","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f436cc7fddabfddf27fba6f7d3ecb065e0c4351e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":66,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":66,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a897f159c8db292fb23f","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-73656c46fa925e494a7b6183","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"657cbaf918c8cfef54e4ec1449872d08f5ec0818","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":825,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":825,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-539ef17134aed8dc52e8","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-865ea429f2d3d0eb7f158169","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"99748794df887461cdf8a4e6879eb0fee59531ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":802,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":802,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2c2212160145e95ff171","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7245b10588d76863511c8f26","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"73.077","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1e1f70985650e54ed44ad05fa6f610d33d8a26bc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":89,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":89,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5e8098917798e44f98d2","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3af822d5cbc9d359160bed8a","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3281d42c32ea7365998ae0714129302afd17c5e1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":112,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":112,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d59b3e91675085a3dc1c","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b87931be996ee1c1f652cc7a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.077","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f41ff69aba3133f38af03d2fbab750bfbc36900a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":135,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":135,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c67fe06692164cdda2cd","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1559be248c4e16da0d4f9b4a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"86.538","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f41ff69aba3133f38af03d2fbab750bfbc36900a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":848,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":848,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-415c902279420e2e4a1e","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2801a29a8ad68911c5bcf350","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e59205c1238c13da45b3cbc61fce5299aeb41c0f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1055,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1055,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-29e40bd7ba6c21211547","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-445440de9911fe31201a5a49","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"37b61574acd959ede96daf49bd559479af3c4b70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":158,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":158,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1fc4612791b35402b6bc","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6ff76ce20cfcd321298ba3c3","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.615","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"62f022d77c225a3c47744043b8aea3386d6d0502","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":940,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":940,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-508c574f10f4e9bb9ea5","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4b2eee03adabdd21e05f0148","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20ee1346c4e6b013a4158ee35dd7786358b7f195","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":181,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":181,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-903a62f747434e561c2e","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9e56e6bdab7814ba6666b86a","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ada8fc5ca476f2471c2ebf6b60a7705700405f10","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":204,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":204,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-977630c4368e810c9e32","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-855301fb50d157d8d836c471","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8cb94fd643598a203c0762c23ca36b6a3fd4846","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":756,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":756,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d3291564607f8e388b75","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c8f2d2337acc7b852326c273","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b77ad13aafb2d31e03168819b2a3393faad8f6cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":779,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":779,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cec98f8cec50aea97a8a","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d64f78abab51bf850381e38d","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.692","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d79cbf1c6e26f00ec6dc90649844b55d0942bb27","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":986,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":986,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d18ca6f3561b298a8755","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ad689680aa6bdd09279419bf","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6d3a7a7ebacd93ed45d745564cc0d7b5cacff457","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":227,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":227,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5933b5dda5b9ee6045e5","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-93522971f29e776d8b78ef9e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"88.462","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5137e8d2ff4b6fb141152a5afa705ade12cc3ff2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1078,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1078,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.462,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-620f1a9a47c35cb68c42","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5e72a3a063cc8b78c0338913","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"73.077","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b18e56c19354793f2d1ddebb0c7686671088ca09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1101,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1101,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3262d1aaf9b1294f80a9","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ec61b2866e24e12b867e46be","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e1b588a3efae252cf59c75fa0602275ee5f721a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":250,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":250,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1ec332b2ce5a9bc49f30","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-72a55783968d34a41d14542c","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f3ffceb032f94464bbbb06084499b8796e4d6561","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":273,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":273,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e33e69378e035bb2c6d3","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a50170a7a9b917c8c95c70ed","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"76.923","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ee23828bd05128e71f0c3dd8ea95ce183d8535c1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":296,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":296,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.923,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-01f58f70f43a1aed5644","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1ae832b349e1787452255e97","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e10203d14f41691f16ee7d4f148e8ab5341b4a4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":319,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":319,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f0a84aac520041ccd7b8","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a705eaa8a9ad0ad2e64aa1d7","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.462","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"95f1acd9889028fbbbb0c6fe9e7fcfd46201b133","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":342,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":342,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.462,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-82ae653854bb0bd61c5d","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-593438d0c382ebd0112002fd","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.615","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3011686d3572a1766cc1d44c5dfa78809f2f6584","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":365,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":365,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-038242c5f22e3b8bb44b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-296d0dbbc1ff7cb8f9314ab0","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"73.077","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"96344c6a455bfea108b46bda993e67090a8fa4c7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":641,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":641,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.077,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d82c932695d2390b3a8e","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-04fb3124c4ccc6732da8ca75","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.615","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7aa6ceb915a85d567cc47d10bfac8e106264754c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":595,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":595,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-26872e7c56340b0bab8e","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5c22bb0ed1c3260e10091d35","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.538","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e95acf95a998c278ac324e9525480a61bd6d510","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":618,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":618,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-87e81e349a5f2a1e3d07","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e6df08b1b7a11fc40fd2b8c4","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.692","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4bf8545e33829afbe175fc1ec4839896f88d2b23","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":411,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":411,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6a12accde9c548fc0eec","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-816a144d6b5cc472aa767256","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"86.538","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"443ea45ffc9c9323492646ada691f66e84788ab1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":963,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":963,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.538,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2b58346be6d1114dcd85","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-19c4e7805ceca3b99435999a","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"75.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"967151516e8454f73101a09177ae295dc95e7588","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":457,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":457,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a31fcea564d0edd3945f","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ff9e912b0f46df7be4a5ae0d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"69.231","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f67331aef88a23d37cd58355af09b926a59bca3d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":480,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":480,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.231,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-6cfa1450add882cbbfbc","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c9fc3ffdb22cea2010c6965f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"82.692","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2a59b0a1000ef28ce445959f105cd2e38145308a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":710,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":710,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.692,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b718b249ad2f457c03ba","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6d39cdcca601c68257c54772","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.923","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"651554e1fc92d5e4155168bf4d17a84db0eaeaa0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":503,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":503,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.923,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-56dea86e23344b6995b0","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cc0465122d95b4e1354d45f9","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"65.385","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"acade7c7c08cdd358e81c14386d6bae922a87961","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":526,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":526,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.385,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0b70a049acbfc1e594af","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-54f3edde13df683a98c52c0a","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"67.308","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"74ee1a2a3f8e9cb33f25d45fbf329df5a097ddf6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":549,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":549,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.308,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fedeae95456e32dab0c8","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d8f5998c0c0882088773d392","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed9fea4c623728b57c17b0fbeba27bf7ba2b43d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":572,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":572,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-191ae0d07f678afd1966","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-daab61522fe074a70725096f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"59.615","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8a33f1d43ebbf82fcc0e0bf01622865828302329","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1009,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1009,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.615,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a7f47fc8b99b7f597038","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ab6104df2a3042fd04542761","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.769","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9f77cdfccf529a6b2bc8cfb1b4215dcc4d91137f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1124,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1124,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.769,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-theory-of-mind","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2b808d806ddf8502c206","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-31dad745c7a5d044d7026745","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"theory_of_mind"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"78.846","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"60365b7f645a65548846e3c93787f1c303ffd7be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":871,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=theory_of_mind","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":871,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"theory_of_mind"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.846,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-82551276d6a64a204567","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fc02427935d08c13e2389924","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"53.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"87ae3be67d69fade425a64293caec42d9f1c7d5a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":665,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":665,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":53.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a72362f3a56d8c3422c8","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-67a2855da4a5607dbd69b623","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"20.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9ae6b49bf1eaf1f7ae7ebbe99c5f7222630b83d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":21,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":21,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b0c1d8af96bb78556c1d","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-be5d7b27bb4b3b033e809eb8","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"33.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cf7cafc34c252ac37ff5e7f91743571efa22c017","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":44,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":44,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":33.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a2e7175c494c56eaf7d9","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-08f8c9a73eff2ddbe9c16c3a","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"43.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d246052a2c0c409343fda59828649ec9663c2d3f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":67,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":67,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-cbba95a54a3735e9a93e","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fa9e2d144b0661b8d88798de","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"33.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a5c1adca4e9bc39248bab3df0565b21b9cb3b506","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":826,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":826,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":33.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9c1468ab98fd9f825e9f","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e1c9a24ddc895cbf9e863f8c","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"43.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f66f6081c2be2a2236e96d4d9905701023558870","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":803,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":803,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c13444064079232a29da","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ed8ea7c38242c23fea6f1a41","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25db9e9229d22f1765763db4c4ac8d1005e6e82d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":90,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":90,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a8346f7d188450a038ac","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-79a674e10f865659bbef67be","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1ad35fbaedc60f937ef79dbb5da58669d7a2971","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":113,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":113,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-077b5af4fb7c1b423b15","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0bf06bde22c000d4a273654a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21fe7fc04e57808cd5bd49d1ea6690d583b900e2","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":136,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":136,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5a2ad2b1831db4d87825","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7cae398d0292d08a6255babd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"26.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21fe7fc04e57808cd5bd49d1ea6690d583b900e2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":849,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":849,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0827f0e28763d5476068","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b56f9625d448e7ab10110812","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"56.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90d14facb1c0156567360a8c3a083afca849476e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1056,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1056,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c443fd8d573a906ba210","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0734d4004b86609256bf0941","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90b4c6a4b82902ee5dd84b222e52fe244b0034a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":159,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":159,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-23d7a4b394fe1303abdc","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ae6ec69f758f692aa6e06fea","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"46.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4fa5cbae39570c183b6e3ca83e6ff167f1a37f70","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":941,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":941,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4dd5bdc64468546e78bd","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-329aa6e9c34e18049be50b04","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa7018c175f431ad5fed44ee4a6cbe5af71d3d6e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":182,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":182,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-674867a2ec560e49101d","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2a994f489b884b2a7c6d1921","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"33.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f5fb4eca317b2d98109c1e0edc8b9589c13e0f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":205,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":205,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":33.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1a92a37769343a6b9e07","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7e6d415a0dead0f4b0b22b40","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"26.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f6f963aeb8eaecf6cd2f0e41243ba15a8cf6c63","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":757,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":757,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-186a86dccd448fcbe730","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-08fdf9f677d6f155ac11a80d","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"26.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83e710980ee9264420e41d813a96901ec726fd93","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":780,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":780,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ed0f002492bd3c8986a7","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-89937b8494ac3506146e59a1","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"46.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3919e5db7d58f47d0778dcd764ae0ef62fa76d7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":987,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":987,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-df0c5ebaf579328ca2ed","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e57af9a0379e696caf463ab3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"36.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5455aea036f945262ced9d2ab31f24319d96e34","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":228,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":228,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":36.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fde4af5806616294f3e9","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-16e8192b4d460807a2df7c19","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4e653efc3abeeb3968ea2e2c0fc819fbf982273","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1079,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1079,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e4b5599f8a0ebc1c6fff","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-db1d2855cc51cdf6a7f715d5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"46.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a425ea0446d09de39f28cc58dfac1dcdbd7656b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1102,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1102,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3c3367783a503500633c","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4c986b5270ad1e20785b4faf","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"36.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"524d9599c6741cc0605c790d87d0f4a34d87288c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":251,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":251,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":36.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3ef6d14995a4f6dcdfe3","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8a42a17bf4e2a680b6557c83","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"20.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7a232998ebaea17f87e3a3ebc75f1e705b43af75","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":274,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":274,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-04f000faf35644321a73","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-79b1878a978b88249bbec360","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"20.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dbb7be1a5f40ca504ff45a9416bf7c05150eabdd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":297,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":297,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fb69b6a2861de237807b","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-234ce01e9dcb23191bc3f157","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"26.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"52041961f585b682697df660a428f7e204bde3be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":320,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":320,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-175e241718899e228aa5","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-afac7d38b9dd5c5c6d9f6fab","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"43.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23c417960bf9e8b453ec24a60caa13213177a4b2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":343,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":343,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-97fb123513304e6c8570","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f3daed7b5bcf25136aa3744e","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"43.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"888a1d57a48b4af4c3a55ac7f8b5f26bb591a626","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":366,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":366,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dbe95f24844ac1486ef7","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-56ef582db0c627c6a8a2611e","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"36.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eac8d34f797e5806cc97ddf9e3887589931c503a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":642,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":642,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":36.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9695ccb84ff631a36735","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0b9d137e095629da92b6b7ce","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6923613c65c655e249f107c135e0cdaa4ea2e88c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":596,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":596,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e423e35805fc537b4225","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b670c095e6b63e2f7ff2f139","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"46.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eccc7936dbc2ee7f2241c6a0c9a968fad3ab3a64","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":619,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":619,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-efb32f9bfece6711e6a8","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a11b1ad9350cf20fe7c9e5d3","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"36.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa9384cf84f87238419bc619157c2c0306b502a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":412,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":412,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":36.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d78625a3bcd64893acf4","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-763adcad8c7248c25fd41a3f","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"43.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eb3425fbdf8278723ff040ad49e53f7ec78f0392","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":964,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":964,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b238db50804548067b14","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-69213a2395cb791da96da84a","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"26.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"617ce302c8894b985de63596db262e728bfc88da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":458,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":458,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-78db9345447eceb4831a","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b52b409391e51eb2399bbc5b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c97b135a0901270548f1e24c917f561b4514b30d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":481,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":481,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-436fc5dbd7e0f18430a1","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4a700006a72b3bfd113b15a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"53.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f371d6e6decfbfa97e9a5ac59eeffac24786eec","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":711,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":711,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":53.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a98877308e06b87e4922","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1d44fed410044847a50cf3cd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a547f0e42f1ed1e303197b7ed910d3912dbc3218","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":504,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":504,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a34d857bff13f8b8b8c4","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-54b6be2e3df15b1da5a882f3","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"23.333","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"885d2eab071243baa257dd5b5f1bc4fa00b1e4e5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":527,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":527,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.333,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b3ef51b55e776444cdb2","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-426a57197c534ac619824915","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"20.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3ef6f87d75a15dd85ad978999a53de4660adfc7f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":550,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":550,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9bcd6d9c9b9a17e5e70f","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dcc50f42a94e3ffcd1ab4044","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"26.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"188a34df9a6908d17e70481fc1b0716c4c7f8c13","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":573,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":573,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":26.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-7cedc7d044c1e1137777","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-835221793fb2e9bc4e4415ea","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"50.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"75171911f6826f05aa5b0c7071330b8f2185ac9d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1010,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1010,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e2977a5e6476930ebe3a","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-47d3f0c1fad0588a3a13c069","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"46.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"de45145b44f4f5132b1bbd4751e3d69f16163fb9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1125,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1125,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":46.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typescript","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-37299bd27b018195d386","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cfe4fae90143c7aa1177994e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typescript"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"56.667","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c556c3debf472c849bb255a24481628e8bc953f3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":872,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=typescript","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":872,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typescript"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.667,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-03f3238e960b6a4e8486","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d8f53fc811e93c6987ea382c","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0f720fd9e096d58aff919748e5f70cdc1690ec04","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":666,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":666,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2dcc442b7915143f2893","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d50485487a6a6dcf4b15a794","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d90bb49d3d1d6b7ef3cfe586bc85140eca14a909","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":22,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":22,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2aa32f59c22490184675","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-47c87d926e7895f420510e8e","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6fcbab941ae9004ca4556db0b6b29b7db2f80a22","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":45,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":45,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-00ce693066e11fb5fce0","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c70c318153fa6c3abf4bef32","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7381cf21d3fccee73b6e5b398dfd7cce22cc7007","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":68,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":68,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0f7b9c119d748835eae1","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2738de63197644e0bb68994f","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df148986f9a574bdb12ec5f58ad85ff36e025897","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":827,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":827,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5e00f956e055a830f15e","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-15b58897be3e47474583dc47","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"92.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59e1430fbc83a27f65708baec35d40424395f9ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":804,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":804,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2d22d36184a3308fe961","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bce0d49d4c65ca59cc7d65c2","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"72.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e0b3ba1c63143a55b1733333b092b58c26b5963","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":91,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":91,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b197fbddf53dab7d57c6","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a674a6b6030c4aafd8ca934","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f0834f2ff9209de4b9a094fc5445721f8c25e630","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":114,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":114,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-059f5e3c95a52eebe982","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b9abc6b994abd8a9b4c79201","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"655c001719bbc94e4c80a94a0b0c7e0f7c893baa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":137,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":137,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-dce22158e7c06648fde6","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b41a95ab2a7f17e13c23e25a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"655c001719bbc94e4c80a94a0b0c7e0f7c893baa","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":850,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":850,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-295dd380601eeb7d5dfa","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-52c3f09af03a649205924d92","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59905f923e816875c3d1d18e42318aa74b08167a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1057,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1057,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5ca2a53ad70b466ceff1","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-791effc36acce69762e7dd7b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"772696fe1cb382da186bf99465ba6e17e4af4b4b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":160,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":160,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d3a50dfedea642f1996e","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a460f0232f7ab55928cc5b3a","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c8f6a5b67dea17232e1cbac00864db4851af229","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":942,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":942,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-256b562b89ef75c0e1f2","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-7aa926224dabc164f78f3e18","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83e0c9ac5fa0f8dafd3c61d3e58f9f5c0be502da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":183,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":183,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a9bc058c8a94b7e68c94","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-899d6b771191bf2a6f69fa18","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"86.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"222e334c3ac483f6646e5279b2c70a08704a6c0f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":206,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":206,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b7555f5f01c3b499951d","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-04bf2d6a075b075d80faed2e","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6be574865d343d048369bd154ad7aa7385e92c47","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":758,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":758,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2deb2b0aef7da7c4b319","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ee322baf2c8329251839a692","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5ba71e9c53b27dee73b01e5a8c0765611faadcc7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":781,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":781,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c60052dbd12334f437bc","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d05134c19aab0e640b373acf","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"90.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8576224d51b36b3f85b1c9083771a7c3b6939cb4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":988,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":988,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":90.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fe4ece2128e1e7f66517","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5a345155f80d24a5854f9027","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"74.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b4e3bdfa175c623bd152f2e421a3e392ffde268","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":229,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":229,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ae596376305f02700250","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1a1f1c771f9bed3ed634cfd5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"53c6c2c323bd8b4db9325f4fc056f6e4cb5bbf94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1080,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1080,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-eb0ef87a6596791070c8","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2d409214e2ddeaf2ce171539","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e6feef171445f80fb71e2235cc3dd7486345f598","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1103,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1103,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-c31c025bfb630534ae95","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6cb8e08cea50c71716586d20","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"873a092918361df14b5e0ca632fe1f8e2fecf092","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":252,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":252,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f377fa18a09eca5ff2e7","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-aafbfc140104e72df84ca74f","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bb5de9625448a35026f0725b28e730762a41796f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":275,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":275,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a181cc4aba7c7397dde7","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5b1c8cbd61a6f9661ad3e851","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7f84de78e291b3b2162c28855478af4e8b90cf09","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":298,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":298,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2d9b3443bac74e1dedf2","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6a6fdcd369cf0c9671f8e209","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"54.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6f9461f9c313a9f66913d3f52a700427eff8dfdd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":321,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":321,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":54.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-560b1df07cc609790ccd","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-5cd11067c87eb2e9837d6582","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0853255b5da6216b15a44cce20581ad4f80afaee","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":344,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":344,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-446570f367f77634bffa","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c6c611ad7e9267155c7c691f","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c0ea610c078841614409c01f68683176a665d8e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":367,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":367,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fa81595a99a9065905c4","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4cb45af04faee8628144f069","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6da400208da4211ad4ae79534a1ea31aae7714c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":643,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":643,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-770c43c199ce449d0288","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e22ffd47e073ac5f97f6ab71","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7ad7698496ae0aacf1260af6efa72ed64d9bd30a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":597,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":597,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a09727d4bf0d7fdf0de7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ed91b43ea3b9b07a4fb6e2be","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"86.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d4da6f5be4ed5e796e5ae26bca2ed9ac4843477","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":620,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":620,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-643c12eec643533ce133","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a1faa7bf0d920cc2c11c1b02","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a6d83216b20488a1dedaec43d0e951c3bc85c59","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":413,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":413,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-2a793421b57c2b845ef7","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-85de14795e47ac3a5aaad168","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3937a9ab5ef5341f2f31a389e22bec8509f6fbfd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":965,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":965,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1a966d2787acc400e812","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d073ec90334adbb6d95a1d5d","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9fdf06164c29960618192a0b3b764b5adb9c1601","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":459,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":459,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-089d960dd4655e771c48","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b00d1d8369af19b9a0b3a813","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"80.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"90b8808934c714a307c28631c47e7ac9d1c0fa16","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":482,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":482,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1277e2e22263a53fc9b7","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8857681292a45cb3d906acbe","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ef791e3e8c296b9ec7d091313c886d23963c8484","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":712,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":712,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-305645ed111fa7e9046d","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-b03f77954f9964529fc73d22","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab8874b077a79bae20eb77f77d89e46b73e51c5e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":505,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":505,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-aa86ee37e262964b49da","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-de5c0a1ea18b15deb823204e","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b3702d0a8c711f60337d66f69c6ca5ac284681dc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":528,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":528,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1942f5a78dc046601ad0","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-bd1724de55fc47c081f800d3","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"82.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3ffbd2cc57b57b3b694dce1b4ffc246e8bd51be6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":551,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":551,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":82.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f871bf9829c09ce3604c","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-4e83f7ba8a205171173f401b","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4975477f72c7bf46134c513957494f8cf287e69","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":574,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":574,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-5b7ae5de0824875e3d60","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1d832f2cc6e531da8a8aecb7","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"76.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e29528d711dd0334911e6450725169e70b582de9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1011,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1011,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-da970901db29362edae3","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-1944a34e380412f6f5595a8d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"78.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2fb4d0095be5c1f07bf7087ea67fce4a34541363","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1126,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1126,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-typos","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-9608bc8840dc519779b6","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8f635492eb630a8fe710bc46","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"typos"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"86.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"510093f3cbce50bfa96610897ab27ef3c32224a2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":873,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=typos","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":873,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"typos"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-119e49f949116ecf4905","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-841965cd64777394b35b567b","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Fable 5","modelRef":"claude-fable-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f97c4574be3b68fe299787a4f8f54cb13c50b54c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":667,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=28;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":667,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-17fb0324bbd5a4792d4a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8bd85b5cfd4e7628b102b405","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101-thinking-64k-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"77.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9699fcef373902cc321e074cec3983bbcce8fee6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":23,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=0;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":23,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-da4edf9c2ee797bfa53f","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2a70e84b009c037c69abaff9","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6-thinking-auto-high-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9d9cf4d00a30c429d1a9e7815aeb626b38890ecd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":46,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=1;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":46,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-be8dcc5e3211528448bc","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-dbf70e13d7419c3c38318bd9","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8c925712944f8e85dfd16c3210b946323fdcca13","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":69,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=2;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":69,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e32f857005363291c957","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3f9c942f3bc684c257d2de5e","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93ab2000393923a63a6944fcfa44ae4cce291980","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":828,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=35;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":828,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-47fadabe91b02ed80a8d","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-fd93c8de6e4f1e9e56fdaf8b","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Opus 5","modelRef":"claude-opus-5-max-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f52a8bcbd01d46bff55f763b1bf3c7441c41bfda","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":805,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=34;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":805,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0dcba4dadfd69025c397","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-68f3a7a000a226b3e4448b8f","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6-thinking-auto-medium-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cad5c96ac8b76a022a053aa053644c0b0596b781","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":92,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=3;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":92,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-daf91f0b52ed028fde70","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3135999f25fcff2a93ad667b","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5-xhigh-effort","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"88.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3578a86b79c291359272ad9fe285776218f91472","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":115,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=4;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":115,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-46484f2b3c814c6f8d26","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-33f94d65e0c30274ac386206","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"49.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57ada4e9b8210a5dca2c120da8052f8b9648057f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":138,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=5;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":138,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-734251730752390b448b","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-17d06b9e90ef5d2afd6b165c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57ada4e9b8210a5dca2c120da8052f8b9648057f","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":851,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=36;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":851,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d48a3288c4c2a626b6b0","canonicalModelId":"deepseek/v4-flash-vision-exp@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-54627c6891c29bca8fc1903d","mappingCandidates":["deepseek/v4-flash-vision-exp@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash Vision Exp","modelRef":"deepseek-v4-flash-vision-exp","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"98.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c746c42d66e7cdfe04ae00e79ca50779543cf97c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1058,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=45;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1058,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-4867698d84de73431ee4","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8ebd10a388f0065f5bdfe904","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"92.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c72b0fbf99bdc6e8289cc910b32f5ca1c92bfd28","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":161,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=6;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":161,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":92.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-d3f3c421e36f8da2864d","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-d213716b45763cb8e9831f0a","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"96.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f0bf9417a33bbe8925fc223b9510bac9840190de","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":943,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=40;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":943,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-51afb47e30a1c32d4978","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-33bab6bce4abd21ba5cb2b9a","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"85.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"27ce4610a6fbed9d026652f567a472cb52420e8a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":184,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=7;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":184,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-683b9a6fd0465f2bae96","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f7dba1b4b412213bf11c28d8","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"77.25","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0dd9764a7a5f4b8c031e08a24a9945d9ce297781","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":207,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=8;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":207,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.25,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0c24cce6dbcebdb51121","canonicalModelId":"google/gemini-3.5-flash-lite@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ce9324cfe01e37c93097434e","mappingCandidates":["google/gemini-3.5-flash-lite@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.5 Flash-Lite.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.5 Flash-Lite","modelRef":"gemini-3.5-flash-lite-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"32.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"32654d635c9770b0f07edb97ef89cb4ccc337729","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":759,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=32;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":759,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":32.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e889be2b5fab3dc88c24","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-eaa77d100747ec8b25d52327","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"85.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cdc016d103ab89fd0a16acbac7c44cfd0d6f2fb0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":782,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=33;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":782,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-b5f10efd293aa3751c21","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2d662589e16219cbc26d4189","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"LiveBench high presentation of Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"LiveBench high presentation of Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adda06527deb3bfdbddfea81cd7962e23fe2964a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":989,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=42;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":989,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8a1566568713d87a58a7","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-56d6d2d8a5da76bf3c8ae7d3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"glm-5.2","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"65.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8adc253f5cee70adea328ca42270e95a092b249e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":230,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=9;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":230,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-55f72f75b2e1268f647f","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-17aa327266136a7ff121d680","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3","modelRef":"glm-5.3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"88.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb12884318738c2645015bd3bb2c083bb2aa4be8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1081,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=46;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1081,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-11cf9951fa62b2ffe5b2","canonicalModelId":"zai/glm-5.3-flash@2026-08-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f3822ebbbea2988c2fb8107d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.3 Flash","modelRef":"glm-5.3-flash","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"79.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7668ba15729b52c6139cc7dc0b45f0a1118801cd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1104,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=47;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1104,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-a2e407d1f08a8f144cc2","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-e6b81542db577c8d0d18c155","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11-high","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"84.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6027f3412f2ce24132146fc40ab5ab26da6b513e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":253,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=10;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":253,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":84.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1f24af682afc7b7f55ae","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-cb1fff47d9347127c1ffaa87","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.2 Codex","modelRef":"gpt-5.2-codex","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"70.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"92689281f8fa2441196c2a0662f230b2eb45a015","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":276,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=11;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":276,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-71eb615f6ee744103c46","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-9251b93dcafbfd59bcff7265","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"60.372","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0a9ebc57f247e2de5cde30eed0dca7b290f3102d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":299,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=12;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":299,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":60.372,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-31589b748e23f893269a","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-f277b80a63a9525a08b398c1","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","registry":"model_aliases.json"},"mappingNote":"LiveBench xhigh presentation of GPT-5.4 nano; effort stays in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"97.619","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5178ab99693b04181a4ed185519e506083df0798","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":322,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=13;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":322,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":97.619,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-fe5d0f7a36406924c608","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-2bb6b39e731891ee1fd6a455","mappingCandidates":["openai/gpt-5.4@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.4","modelRef":"gpt-5.4-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"98.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"acba3c9a22ee0fe25f677700b7fb47aa450387f6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":345,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=14;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":345,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8e93fea48cc8d342f968","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-3a1e059bb4fbdbb12b585254","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.5","modelRef":"gpt-5.5-xhigh","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94bab274b1134299880da15b7f6e6fd6ccd07222","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":368,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=15;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":368,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-e9898a7fdc417da7b86b","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-593668a075c6b909959a3e8b","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"93.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"15804119c89c4a6787cb04947506e21f765898be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":644,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=27;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":644,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":93.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-24ff1bb2001df1fc8678","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-878e675a42f00d5dd1fbb888","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed8dd4aed54c4f8c79f2ed2ce1ffdcbd90b12036","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":598,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=25;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":598,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-34edbb8e46258c59ab91","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-a31b098197e09e3f1093feb0","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5289278812d3a4ceb07dd63610d9c0543c70ee94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":621,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=26;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":621,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-1f5fafd747e682a9f6de","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-c3b3d2e0885f3ad52608f8f7","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.5","modelRef":"grok-4.5","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"94.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9315e8b5a86016d597fd5fb03efa031562c80a25","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":414,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=17;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":414,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":94.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-8d8331b3a14c7ddd61d3","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6833f9f2f08b162398dbc02a","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"LiveBench base spelling of Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"LiveBench base spelling of Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Grok 4.6","modelRef":"grok-4.6","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"95.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d66934a803239aa40b9cad76843ffb9bc24313ff","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":966,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=41;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":966,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":95.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-0a7efdeae238cd8ff392","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-945c4c69a7ce6eeb1e8107c1","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6-thinking","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","source_version_label_without_id"],"rank":null,"rawValue":"78.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f53983055b32bad6f8f8a47901b8fea9b950b17","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":460,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=19;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":460,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-758823c96213b867f880","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-6e9771376ec70307a7a657d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"347e3836adbf1b6c58fa7e9765ad4c6a24931441","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":483,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=20;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":483,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-f7bdb6a0a6672dcfab90","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-530eaa0d53ca460a76ee6c59","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K3","modelRef":"kimi-k3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"44c02c809b9edb0f79166703f147ea7228c017c0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":713,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=30;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":713,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ab9ac77353dade06b566","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-47df159992b363c1a154a2d8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"minimax-m3","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"61.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f18e24eb18792417ee58934a21333baf59b78f2c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":506,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=21;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":506,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-52546dbe5775a7c26839","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-938e833e1abc169734c5b815","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"qwen3.6-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"53.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"421dc9393208f8d4aa1d974e5f5e2583373f14d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":529,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=22;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":529,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":53.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-ed91ed22cdc1fd975931","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-8514f0d2c525457b23ef7623","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"68.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"575e4d91365008d812913588c7eb8da201a4b112","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":552,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=23;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":552,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-28356b7ea8da87e8ab69","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-0be0c43c0040c41aca60e8f7","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["curated_model_alias","source_version_label_without_id"],"rank":null,"rawValue":"74.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b693c4e224a5a80b4069bfeae613d7c012599644","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":575,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=24;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":575,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-3580b99ee92914e3a1e1","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-ec56896f17d6b5814273fe75","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"qwen3.8-27b","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"96.5","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"922d850c7ab6eac9fe85761bbc8e96e2b60ac95e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1012,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=43;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1012,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":96.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-464de5b8714f3d92537e","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-51eaed25b9abc8279b2d8e5e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"qwen3.8-flash-next","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"98.75","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c2768951e4e01a3c85d731621f95df9c05154b6c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":1127,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=48;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":1127,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":98.75,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"livebench-zebra-puzzle","benchmarkName":null,"benchmarkVersion":"2026_06_25","benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"source_hint","candidateId":"cand-livebench-official-94422ba4e4d1ecf1354b","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","harness":"livebench-official-table","harnessId":"livebench-official-table","httpStatus":200,"id":"pub-161a2f8e6e6cf7a52617f012","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max","observedAt":"2026_06_25","parserVersion":"0.1.0","payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","protocol":{"harness":"livebench-official-table","release_date":"2026_06_25","task":"zebra_puzzle"},"publishedAt":null,"qualityFlags":["source_version_label_without_id"],"rank":null,"rawValue":"100.0","retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"385e00fc78e5c1182c4b15529d4e5a98d0a0c8b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"livebench-official/candidates.jsonl","line":874,"payloadSha256":"d5b2c8331e24dc8662241e174033056037f05b642347b3453c24781a72578b2b","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":"https://api.github.com/repos/LiveBench/new-livebench/git/trees/main?recursive=1","sourceId":"livebench-official","sourceLabel":"LiveBench · official dated release table","sourceLocator":"path=public/table_2026_06_25.csv;row=37;column=zebra_puzzle","sourceMetricId":"score","sourcePageUrl":"https://github.com/LiveBench/new-livebench","sourceRow":{"artifact":"livebench-official/candidates.jsonl","line":874,"metadata":{"release_date":"2026_06_25","source_status":"published_release","task":"zebra_puzzle"},"sourceStatus":"published_release"},"sourceSubjectType":null,"sourceUrl":"https://raw.githubusercontent.com/LiveBench/new-livebench/main/public/table_2026_06_25.csv","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":100.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-0800d3bbef22909b8e8e","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"AIBuildAI","harnessId":"AIBuildAI","httpStatus":200,"id":"pub-d18e3d06784c2ced3b8ee1d2","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"Claude Opus 4.6","modelRef":"Claude-Opus-4.6","observedAt":"2026-03-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"AIBuildAI","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"63.11 ± 0.44","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fa818a11d086c87a8de14691a65996cc0af14277","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":8,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=11;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":8,"metadata":{"agent":"AIBuildAI","all_columns":{"Agent":"[AIBuildAI](https://github.com/aibuildai/AI-Build-AI)","All (%)":"63.11 ± 0.44","Date":"2026-03-06","Grading Reports Available":"✓","High (%)":"46.67 ± 0.00","LLM(s) used":"Claude-Opus-4.6","Low == Lite (%)":"77.27 ± 0.00","Medium (%)":"61.40 ± 0.88","Running Time (hours)":"24","Source Code Available":"X"},"date":"2026-03-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":63.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-761e400a09dea00188be","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"AIBuildAI","harnessId":"AIBuildAI","httpStatus":200,"id":"pub-1b7897cc32e211442b53e573","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"Claude Opus 4.6","modelRef":"Claude-Opus-4.6","observedAt":"2026-03-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"AIBuildAI","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"46.67 ± 0.00","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5a2061c5f92739265d80399c0d7abdbc255abccf","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":7,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=11;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":7,"metadata":{"agent":"AIBuildAI","all_columns":{"Agent":"[AIBuildAI](https://github.com/aibuildai/AI-Build-AI)","All (%)":"63.11 ± 0.44","Date":"2026-03-06","Grading Reports Available":"✓","High (%)":"46.67 ± 0.00","LLM(s) used":"Claude-Opus-4.6","Low == Lite (%)":"77.27 ± 0.00","Medium (%)":"61.40 ± 0.88","Running Time (hours)":"24","Source Code Available":"X"},"date":"2026-03-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":46.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-34b364ed5da82510fcd2","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"AIBuildAI","harnessId":"AIBuildAI","httpStatus":200,"id":"pub-922f012b1b271363e6c68bdf","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"Claude Opus 4.6","modelRef":"Claude-Opus-4.6","observedAt":"2026-03-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"AIBuildAI","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"77.27 ± 0.00","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"028fcbe80dc2be8fd70d4a1aceb16b2e5dfe115d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":5,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=11;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":5,"metadata":{"agent":"AIBuildAI","all_columns":{"Agent":"[AIBuildAI](https://github.com/aibuildai/AI-Build-AI)","All (%)":"63.11 ± 0.44","Date":"2026-03-06","Grading Reports Available":"✓","High (%)":"46.67 ± 0.00","LLM(s) used":"Claude-Opus-4.6","Low == Lite (%)":"77.27 ± 0.00","Medium (%)":"61.40 ± 0.88","Running Time (hours)":"24","Source Code Available":"X"},"date":"2026-03-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":77.27,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-123d9a67d3c8ab28fda2","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"AIBuildAI","harnessId":"AIBuildAI","httpStatus":200,"id":"pub-e8888442506efed604e62065","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"Claude Opus 4.6","modelRef":"Claude-Opus-4.6","observedAt":"2026-03-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"AIBuildAI","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"61.40 ± 0.88","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4ccc30a56da25a9dd46a34a3dea894e900666f67","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":6,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=11;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":6,"metadata":{"agent":"AIBuildAI","all_columns":{"Agent":"[AIBuildAI](https://github.com/aibuildai/AI-Build-AI)","All (%)":"63.11 ± 0.44","Date":"2026-03-06","Grading Reports Available":"✓","High (%)":"46.67 ± 0.00","LLM(s) used":"Claude-Opus-4.6","Low == Lite (%)":"77.27 ± 0.00","Medium (%)":"61.40 ± 0.88","Running Time (hours)":"24","Source Code Available":"X"},"date":"2026-03-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":61.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-1b5f87b8e896135759ab","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent","harnessId":"Famou-Agent","httpStatus":200,"id":"pub-f7e2e5cfc89d3f4ab5b6d38f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-10-10","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"43.56 ± 0.89","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a254bda834aa12f38d10bf751affa5f466f52577","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":52,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=22;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":52,"metadata":{"agent":"Famou-Agent","all_columns":{"Agent":"[Famou-Agent](https://github.com/baidubce/FM-Agent)","All (%)":"43.56 ± 0.89","Date":"2025-10-10","Grading Reports Available":"✓","High (%)":"33.33 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"62.12 ± 1.52","Medium (%)":"36.84 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-10"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent"],"subjectType":"system","unit":"percent","value":43.56,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-d9ca711c41cb66a954a8","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent 2.0","harnessId":"Famou-Agent 2.0","httpStatus":200,"id":"pub-d844f54714fc905979ad4f3c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-12-27","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent 2.0","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"59.56 ± 0.89","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a254bda834aa12f38d10bf751affa5f466f52577","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":24,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=15;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":24,"metadata":{"agent":"Famou-Agent 2.0","all_columns":{"Agent":"[Famou-Agent 2.0](https://github.com/baidubce/FM-Agent)","All (%)":"59.56 ± 0.89","Date":"2025-12-27","Grading Reports Available":"✓","High (%)":"40.00 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"75.76 ± 1.52","Medium (%)":"57.89 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-12-27"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent 2.0"],"subjectType":"system","unit":"percent","value":59.56,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-9e19ee541199c4a6f0e2","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.0","harnessId":"CAIR MLE-STAR-Pro-1.0","httpStatus":200,"id":"pub-8bb1a3c8386b53974ab612a5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.0","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"38.67 ± 0.77","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a254bda834aa12f38d10bf751affa5f466f52577","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":60,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=24;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":60,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.0","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.0","All (%)":"38.67 ± 0.77","Date":"2025-11-03","Grading Reports Available":"✓","High (%)":"31.11 ± 2.22","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"66.67 ± 1.52","Medium (%)":"25.44  ±  0.88","Running Time (hours)":"12","Source Code Available":"X"},"date":"2025-11-03"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":38.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-321204549e2991bc8418","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.5","harnessId":"CAIR MLE-STAR-Pro-1.5","httpStatus":200,"id":"pub-125b7b9b6a8ca1a845b45925","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-25","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.5","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"44.00 ± 1.33","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a254bda834aa12f38d10bf751affa5f466f52577","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":48,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=21;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":48,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.5","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.5","All (%)":"44.00 ± 1.33","Date":"2025-11-25","Grading Reports Available":"✓","High (%)":"33.33 ±  0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"34.21 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-11-25"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":44.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-bc7cfbf52a5168321c74","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent","harnessId":"Famou-Agent","httpStatus":200,"id":"pub-82a88f0ed43374a2473f8429","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-10-10","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"33.33 ± 0.00","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8ad9c572078c0cd13b5dad232bd05c744da5d3c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":51,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=22;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":51,"metadata":{"agent":"Famou-Agent","all_columns":{"Agent":"[Famou-Agent](https://github.com/baidubce/FM-Agent)","All (%)":"43.56 ± 0.89","Date":"2025-10-10","Grading Reports Available":"✓","High (%)":"33.33 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"62.12 ± 1.52","Medium (%)":"36.84 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-10"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent"],"subjectType":"system","unit":"percent","value":33.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-590093e7eb8ba7b22cad","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.5","harnessId":"CAIR MLE-STAR-Pro-1.5","httpStatus":200,"id":"pub-801cc75aa5a62bbe9c0f193f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-25","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.5","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"33.33 ±  0.00","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8ad9c572078c0cd13b5dad232bd05c744da5d3c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":47,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=21;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":47,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.5","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.5","All (%)":"44.00 ± 1.33","Date":"2025-11-25","Grading Reports Available":"✓","High (%)":"33.33 ±  0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"34.21 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-11-25"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":33.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-4469d96e7fc8cf495e50","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.0","harnessId":"CAIR MLE-STAR-Pro-1.0","httpStatus":200,"id":"pub-6ec17d8bbbddcfd84c94f9dd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.0","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"31.11 ± 2.22","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8ad9c572078c0cd13b5dad232bd05c744da5d3c","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":59,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=24;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":59,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.0","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.0","All (%)":"38.67 ± 0.77","Date":"2025-11-03","Grading Reports Available":"✓","High (%)":"31.11 ± 2.22","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"66.67 ± 1.52","Medium (%)":"25.44  ±  0.88","Running Time (hours)":"12","Source Code Available":"X"},"date":"2025-11-03"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":31.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-599b93bec59354237543","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent 2.0","harnessId":"Famou-Agent 2.0","httpStatus":200,"id":"pub-1f3dcdbd514c6b963a328cc3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-12-27","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent 2.0","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"40.00 ± 0.00","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8ad9c572078c0cd13b5dad232bd05c744da5d3c","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":23,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=15;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":23,"metadata":{"agent":"Famou-Agent 2.0","all_columns":{"Agent":"[Famou-Agent 2.0](https://github.com/baidubce/FM-Agent)","All (%)":"59.56 ± 0.89","Date":"2025-12-27","Grading Reports Available":"✓","High (%)":"40.00 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"75.76 ± 1.52","Medium (%)":"57.89 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-12-27"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent 2.0"],"subjectType":"system","unit":"percent","value":40.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-decb83f5b8a76b0a912e","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.0","harnessId":"CAIR MLE-STAR-Pro-1.0","httpStatus":200,"id":"pub-ef8626f48171acd78a5bcd43","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.0","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"66.67 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"921401138ee10bd7b184a9f6ea4f3c474a05458d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":57,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=24;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":57,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.0","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.0","All (%)":"38.67 ± 0.77","Date":"2025-11-03","Grading Reports Available":"✓","High (%)":"31.11 ± 2.22","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"66.67 ± 1.52","Medium (%)":"25.44  ±  0.88","Running Time (hours)":"12","Source Code Available":"X"},"date":"2025-11-03"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":66.67,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-5b5c668a4a26feb6d35d","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent 2.0","harnessId":"Famou-Agent 2.0","httpStatus":200,"id":"pub-d3434f176d4cad4a8aeb479b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-12-27","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent 2.0","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"75.76 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"921401138ee10bd7b184a9f6ea4f3c474a05458d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":21,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=15;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":21,"metadata":{"agent":"Famou-Agent 2.0","all_columns":{"Agent":"[Famou-Agent 2.0](https://github.com/baidubce/FM-Agent)","All (%)":"59.56 ± 0.89","Date":"2025-12-27","Grading Reports Available":"✓","High (%)":"40.00 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"75.76 ± 1.52","Medium (%)":"57.89 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-12-27"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent 2.0"],"subjectType":"system","unit":"percent","value":75.76,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-3ef3f1eec5f2420e6304","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.5","harnessId":"CAIR MLE-STAR-Pro-1.5","httpStatus":200,"id":"pub-7d7fa4c99a8c39a3e85c2c99","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-25","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.5","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"68.18 ± 2.62","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"921401138ee10bd7b184a9f6ea4f3c474a05458d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":45,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=21;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":45,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.5","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.5","All (%)":"44.00 ± 1.33","Date":"2025-11-25","Grading Reports Available":"✓","High (%)":"33.33 ±  0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"34.21 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-11-25"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":68.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-d78f8dd0476ca3aebbc8","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent","harnessId":"Famou-Agent","httpStatus":200,"id":"pub-7124d37289a43f3379edd1b0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-10-10","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"62.12 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"921401138ee10bd7b184a9f6ea4f3c474a05458d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":49,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=22;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":49,"metadata":{"agent":"Famou-Agent","all_columns":{"Agent":"[Famou-Agent](https://github.com/baidubce/FM-Agent)","All (%)":"43.56 ± 0.89","Date":"2025-10-10","Grading Reports Available":"✓","High (%)":"33.33 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"62.12 ± 1.52","Medium (%)":"36.84 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-10"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent"],"subjectType":"system","unit":"percent","value":62.12,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-e43b0426be2b102b89ca","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent 2.0","harnessId":"Famou-Agent 2.0","httpStatus":200,"id":"pub-d7940b963c2dc7eb924cb6d5","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-12-27","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent 2.0","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"57.89 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adcfc92d574b069e21e1aa493dee37ea03bed88d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":22,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=15;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":22,"metadata":{"agent":"Famou-Agent 2.0","all_columns":{"Agent":"[Famou-Agent 2.0](https://github.com/baidubce/FM-Agent)","All (%)":"59.56 ± 0.89","Date":"2025-12-27","Grading Reports Available":"✓","High (%)":"40.00 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"75.76 ± 1.52","Medium (%)":"57.89 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-12-27"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent 2.0"],"subjectType":"system","unit":"percent","value":57.89,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-dcfe770ca796b549e838","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.0","harnessId":"CAIR MLE-STAR-Pro-1.0","httpStatus":200,"id":"pub-c1310a5429640281b0b38928","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.0","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"25.44  ±  0.88","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adcfc92d574b069e21e1aa493dee37ea03bed88d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":58,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=24;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":58,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.0","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.0","All (%)":"38.67 ± 0.77","Date":"2025-11-03","Grading Reports Available":"✓","High (%)":"31.11 ± 2.22","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"66.67 ± 1.52","Medium (%)":"25.44  ±  0.88","Running Time (hours)":"12","Source Code Available":"X"},"date":"2025-11-03"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":25.44,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-66d5563aa2b0f1636e57","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"CAIR MLE-STAR-Pro-1.5","harnessId":"CAIR MLE-STAR-Pro-1.5","httpStatus":200,"id":"pub-96278a3cdcfdf2f7b980687e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-11-25","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"CAIR MLE-STAR-Pro-1.5","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"34.21 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adcfc92d574b069e21e1aa493dee37ea03bed88d","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":46,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=21;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":46,"metadata":{"agent":"CAIR MLE-STAR-Pro-1.5","all_columns":{"Agent":"[CAIR](https://research.google/teams/cloud-ai-research/) MLE-STAR-Pro-1.5","All (%)":"44.00 ± 1.33","Date":"2025-11-25","Grading Reports Available":"✓","High (%)":"33.33 ±  0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"34.21 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-11-25"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":34.21,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-c83bff8fe93188ad46ea","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Famou-Agent","harnessId":"Famou-Agent","httpStatus":200,"id":"pub-2d29eadd6b0c537bd2a118a8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"Gemini 2.5 Pro","modelRef":"Gemini-2.5-Pro","observedAt":"2025-10-10","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Famou-Agent","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"36.84 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"adcfc92d574b069e21e1aa493dee37ea03bed88d","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":50,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=22;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":50,"metadata":{"agent":"Famou-Agent","all_columns":{"Agent":"[Famou-Agent](https://github.com/baidubce/FM-Agent)","All (%)":"43.56 ± 0.89","Date":"2025-10-10","Grading Reports Available":"✓","High (%)":"33.33 ± 0.00","LLM(s) used":"Gemini-2.5-Pro","Low == Lite (%)":"62.12 ± 1.52","Medium (%)":"36.84 ± 1.52","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-10"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=famou-agent"],"subjectType":"system","unit":"percent","value":36.84,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-660389f8fcd8e20a13b7","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"R&D-Agent","harnessId":"R&D-Agent","httpStatus":200,"id":"pub-a644d7cb33f963068aab6ae8","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"GPT-5","modelRef":"gpt-5","observedAt":"2025-09-26","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"R&D-Agent","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"35.11 ± 0.44","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"901d4e4ccc0df9a5f190c301c99a58645f7ac236","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":68,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=26;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":68,"metadata":{"agent":"R&D-Agent","all_columns":{"Agent":"[R&D-Agent](https://github.com/microsoft/RD-Agent)","All (%)":"35.11 ± 0.44","Date":"2025-09-26","Grading Reports Available":"✓","High (%)":"22.22 ± 2.22","LLM(s) used":"gpt-5","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"21.05 ± 1.52","Running Time (hours)":"12","Source Code Available":"✓"},"date":"2025-09-26"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=r&d-agent"],"subjectType":"system","unit":"percent","value":35.11,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-cac9669ed95a55b3f7fe","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"R&D-Agent","harnessId":"R&D-Agent","httpStatus":200,"id":"pub-c7a4068bf711334639b2092c","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"GPT-5","modelRef":"gpt-5","observedAt":"2025-09-26","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"R&D-Agent","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"22.22 ± 2.22","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b352eb06cd231710f8010e81c873de885dbbbbe","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":67,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=26;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":67,"metadata":{"agent":"R&D-Agent","all_columns":{"Agent":"[R&D-Agent](https://github.com/microsoft/RD-Agent)","All (%)":"35.11 ± 0.44","Date":"2025-09-26","Grading Reports Available":"✓","High (%)":"22.22 ± 2.22","LLM(s) used":"gpt-5","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"21.05 ± 1.52","Running Time (hours)":"12","Source Code Available":"✓"},"date":"2025-09-26"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=r&d-agent"],"subjectType":"system","unit":"percent","value":22.22,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-6703014a0f4d35fae7fb","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"R&D-Agent","harnessId":"R&D-Agent","httpStatus":200,"id":"pub-672ba632ed6e4a7c00c33717","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"GPT-5","modelRef":"gpt-5","observedAt":"2025-09-26","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"R&D-Agent","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"68.18 ± 2.62","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59a4714057b5ec4fa62ddc2c3a63239abdf8bf63","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":65,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=26;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":65,"metadata":{"agent":"R&D-Agent","all_columns":{"Agent":"[R&D-Agent](https://github.com/microsoft/RD-Agent)","All (%)":"35.11 ± 0.44","Date":"2025-09-26","Grading Reports Available":"✓","High (%)":"22.22 ± 2.22","LLM(s) used":"gpt-5","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"21.05 ± 1.52","Running Time (hours)":"12","Source Code Available":"✓"},"date":"2025-09-26"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=r&d-agent"],"subjectType":"system","unit":"percent","value":68.18,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-643aec0d8ce3e9c3e521","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"R&D-Agent","harnessId":"R&D-Agent","httpStatus":200,"id":"pub-f5cd790fdb9657204c5652ef","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"GPT-5","modelRef":"gpt-5","observedAt":"2025-09-26","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"R&D-Agent","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":null,"rawValue":"21.05 ± 1.52","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b380ec055ee0cf2bfb07e0bb19efffe07328c45","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":66,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=26;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":66,"metadata":{"agent":"R&D-Agent","all_columns":{"Agent":"[R&D-Agent](https://github.com/microsoft/RD-Agent)","All (%)":"35.11 ± 0.44","Date":"2025-09-26","Grading Reports Available":"✓","High (%)":"22.22 ± 2.22","LLM(s) used":"gpt-5","Low == Lite (%)":"68.18 ± 2.62","Medium (%)":"21.05 ± 1.52","Running Time (hours)":"12","Source Code Available":"✓"},"date":"2025-09-26"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=r&d-agent"],"subjectType":"system","unit":"percent","value":21.05,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-1b6b8f6b842cdf65fce6","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Operand ensemble","harnessId":"Operand ensemble","httpStatus":200,"id":"pub-0f3c799f6f718c0ba9064a52","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"all","modelName":"GPT-5","modelRef":"gpt-5 (low verbosity/effort)[^2]","observedAt":"2025-10-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Operand ensemble","split":"all","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"39.56 ± 0.44[^3]","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"901d4e4ccc0df9a5f190c301c99a58645f7ac236","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":56,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=23;column=All (%)","sourceMetricId":"all","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":56,"metadata":{"agent":"Operand ensemble","all_columns":{"Agent":"[Operand](https://operand.com) ensemble","All (%)":"39.56 ± 0.44[^3]","Date":"2025-10-06","Grading Reports Available":"✓","High (%)":"20.00 ± 0.00[^3]","LLM(s) used":"gpt-5 (low verbosity/effort)[^2]","Low == Lite (%)":"63.64 ± 0.00","Medium (%)":"33.33 ± 0.88[^3]","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":39.56,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-618cd7c8c60ddae67020","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Operand ensemble","harnessId":"Operand ensemble","httpStatus":200,"id":"pub-dff066a65d8c6119d3911d67","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"high","modelName":"GPT-5","modelRef":"gpt-5 (low verbosity/effort)[^2]","observedAt":"2025-10-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Operand ensemble","split":"high","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"20.00 ± 0.00[^3]","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b352eb06cd231710f8010e81c873de885dbbbbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":55,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=23;column=High (%)","sourceMetricId":"high","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":55,"metadata":{"agent":"Operand ensemble","all_columns":{"Agent":"[Operand](https://operand.com) ensemble","All (%)":"39.56 ± 0.44[^3]","Date":"2025-10-06","Grading Reports Available":"✓","High (%)":"20.00 ± 0.00[^3]","LLM(s) used":"gpt-5 (low verbosity/effort)[^2]","Low == Lite (%)":"63.64 ± 0.00","Medium (%)":"33.33 ± 0.88[^3]","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":20.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-a73eff967224666751a1","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Operand ensemble","harnessId":"Operand ensemble","httpStatus":200,"id":"pub-3b38c208f8e46aa6c75565a4","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"lite","modelName":"GPT-5","modelRef":"gpt-5 (low verbosity/effort)[^2]","observedAt":"2025-10-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Operand ensemble","split":"lite","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"63.64 ± 0.00","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"59a4714057b5ec4fa62ddc2c3a63239abdf8bf63","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":53,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=23;column=Low == Lite (%)","sourceMetricId":"lite","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":53,"metadata":{"agent":"Operand ensemble","all_columns":{"Agent":"[Operand](https://operand.com) ensemble","All (%)":"39.56 ± 0.44[^3]","Date":"2025-10-06","Grading Reports Available":"✓","High (%)":"20.00 ± 0.00[^3]","LLM(s) used":"gpt-5 (low verbosity/effort)[^2]","Low == Lite (%)":"63.64 ± 0.00","Medium (%)":"33.33 ± 0.88[^3]","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":63.64,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"mle-bench","benchmarkName":"MLE-bench","benchmarkVersion":null,"benchmarkVersionHint":"mle-bench@public","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-mle-bench-3075e44e025a32b5a773","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","harness":"Operand ensemble","harnessId":"Operand ensemble","httpStatus":200,"id":"pub-68fc27bc98b161b978d4b55d","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"MLE-bench footnoted presentation of GPT-5; verbosity and effort settings remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"medium","modelName":"GPT-5","modelRef":"gpt-5 (low verbosity/effort)[^2]","observedAt":"2025-10-06","parserVersion":"0.1.0","payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","protocol":{"harness":"Operand ensemble","split":"medium","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":null,"rawValue":"33.33 ± 0.88[^3]","retrievedAt":"2026-08-27T08:26:35Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9b380ec055ee0cf2bfb07e0bb19efffe07328c45","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-mle-bench/candidates.jsonl","line":54,"payloadSha256":"586bbc19cb2a3e11be61021866ef6436cf4b4d935bb1db088984b567b76595bb","retrievedAt":"2026-08-27T08:26:35Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","sourceId":"src-mle-bench","sourceLabel":"OpenAI · MLE-bench","sourceLocator":"README.md:row=23;column=Medium (%)","sourceMetricId":"medium","sourcePageUrl":"https://github.com/openai/mle-bench/","sourceRow":{"artifact":"src-mle-bench/candidates.jsonl","line":54,"metadata":{"agent":"Operand ensemble","all_columns":{"Agent":"[Operand](https://operand.com) ensemble","All (%)":"39.56 ± 0.44[^3]","Date":"2025-10-06","Grading Reports Available":"✓","High (%)":"20.00 ± 0.00[^3]","LLM(s) used":"gpt-5 (low verbosity/effort)[^2]","Low == Lite (%)":"63.64 ± 0.00","Medium (%)":"33.33 ± 0.88[^3]","Running Time (hours)":"24","Source Code Available":"X"},"date":"2025-10-06"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/openai/mle-bench/main/README.md","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system"],"subjectType":"system","unit":"percent","value":33.33,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-b0e9ff88a81a84b675a7","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/dd/c9/ddc99cec-f5d3-4cc4-9cc5-bb534377b5f6.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-eafa7183bafcda3ee3860ed3","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.1","modelRef":"MiniMaxAI/MiniMax-M2.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":1,"rawValue":88,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"561f9f492f03a56ef0a5aabe5d805a0f95ff4dd8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":1,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=1;row=0","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":1,"metadata":{"author":"MiniMaxAI","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":37,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/dd/c9/ddc99cec-f5d3-4cc4-9cc5-bb534377b5f6.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":88.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-e060715f9ac7f768e1bf","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/c5/4c/c54c4ee8-ff99-4cda-a81f-a2e3a4347fb8.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4fba15c8e5d2de0dadb9f794","mappingCandidates":["minimax/m2.5@2026-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.5","modelRef":"MiniMaxAI/MiniMax-M2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":41,"rawValue":80.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d471bf72db1a81676f76208728fb64147dca63c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":41,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=41;row=40","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":41,"metadata":{"author":"MiniMaxAI","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":228703644928,"pull_request":61,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/c5/4c/c54c4ee8-ff99-4cda-a81f-a2e3a4347fb8.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-094f88293828c4732bb1","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/af/3c/af3c554a-1eb9-4a79-80dd-678c7aebf0aa.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-015f67a892d2eed941ad6155","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":8,"rawValue":86.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a0f8b94aa37830d35ecc4dfcb40396fe910f2c28","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":8,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=8;row=7","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":8,"metadata":{"author":"Qwen","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":24,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/af/3c/af3c554a-1eb9-4a79-80dd-678c7aebf0aa.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-e59bc35719741329e4f0","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/15/df/15df0778-3ca6-4ff9-9b34-41f74dc4ced4.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4ce07650bd2e8098069cbc50","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":12,"rawValue":86.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b099859a7162438132b34645c178fd1b9da44c36","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":12,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=12;row=11","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":12,"metadata":{"author":"Qwen","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":47,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/15/df/15df0778-3ca6-4ff9-9b34-41f74dc4ced4.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-3102676d1f2bff63db82","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d206c11ca3ce520cb7eee073","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":3,"rawValue":87.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"94e4ba965118395a4d46aae514c94deb444ca6c8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":3,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=3;row=2","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":3,"metadata":{"author":"Qwen","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":9,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-b69b63b897117ed0bfdf","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-68e5924c5958bc2ca74eedda","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu-pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":10,"rawValue":86.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0152478fca6048dc2399428c703ac440208f253c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":10,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=10;row=9","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":10,"metadata":{"author":"Qwen","filename":".eval_results/mmlu-pro.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-c9c7b40f1db89ce83eb9","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-c611ab70e0cfe6e45081e58d","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu-pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":14,"rawValue":85.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4f60eb0f95265cc7650d4dc64043d07be254812b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":14,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=14;row=13","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":14,"metadata":{"author":"Qwen","filename":".eval_results/mmlu-pro.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-59bf45c4fbbd164a5e28","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-ce1409fb230d10f29b1439b7","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5-Pro","modelRef":"XiaomiMiMo/MiMo-V2.5-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu-pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":51,"rawValue":68.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"66ee77a901ae1b5e9b1a5ae932ea599aec7fd5d7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":51,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=51;row=50","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":51,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/mmlu-pro.yaml","lower_is_better":false,"num_parameters":1023244718976,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-a9fc9a88c16e042d7e7e","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-R1-0528","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f7415149b45ba0686e1edfe9","mappingCandidates":["deepseek/r1@0528"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek-R1-0528","modelRef":"deepseek-ai/DeepSeek-R1-0528","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu-pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":15,"rawValue":85,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1e11349e804910238190036d182e2f302e126d63","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":15,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=15;row=14","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":15,"metadata":{"author":"deepseek-ai","filename":".eval_results/mmlu-pro.yaml","lower_is_better":false,"num_parameters":684531386000,"pull_request":114,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":85.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-9425058d28f104b666f7","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-fe041462536506b0150f8bf5","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-ai/DeepSeek-V4-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu-pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":9,"rawValue":86.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"50fbc758d808e019e4441fa5654f1883f91e5fb6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":9,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=9;row=8","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":9,"metadata":{"author":"deepseek-ai","filename":".eval_results/mmlu-pro.yaml","lower_is_better":false,"num_parameters":290944616402,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-734ef6e103094b95c92c","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-1bfa1c78530168b18a7bd253","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-ai/DeepSeek-V4-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu-pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":4,"rawValue":87.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3b44f700f877c1c10c73be77c9a668832055b991","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":4,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=4;row=3","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":4,"metadata":{"author":"deepseek-ai","filename":".eval_results/mmlu-pro.yaml","lower_is_better":false,"num_parameters":1598839674782,"pull_request":110,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-e270a9154081ebb0726a","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-57f1e3696a818675154c199c","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":5,"rawValue":87.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"520a353e751a79dba10a68c129f91df51a492f4c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":5,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=5;row=4","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":5,"metadata":{"author":"moonshotai","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":51,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":87.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"mmlu-pro","benchmarkName":"MMLU-Pro","benchmarkVersion":null,"benchmarkVersionHint":"mmlu-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-mmlu-pro-6098db46fa8b087c6d38","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/76/7d/767d7365-4f30-4ac0-9fc4-e0ade16a161f.json","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-82f3d09333ae16685189c70d","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","protocol":{"filename":".eval_results/mmlu_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":86,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"facb42fb7529d39ab16ba80c72e1c781b68173b1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-mmlu-pro/candidates.jsonl","line":13,"payloadSha256":"b918fd3f4cc409b820f49e53234b47d024663c1810bd3b1e6acfcfa3bf220a32","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-mmlu-pro","sourceLabel":"Hugging Face · MMLU-Pro leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","sourceRow":{"artifact":"hf-mmlu-pro/candidates.jsonl","line":13,"metadata":{"author":"zai-org","filename":".eval_results/mmlu_pro.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":80,"source_name":"EvalEval","source_status":"reported","submission_source_url":"https://huggingface.co/datasets/evaleval/EEE_datastore/blob/192329fb7d6b15b7b0936a1a58ae862aa7e8ba24/flat/objects/76/7d/767d7365-4f30-4ac0-9fc4-e0ade16a161f.json"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/TIGER-Lab/MMLU-Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":86.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"osworld","benchmarkName":"OSWorld 2.0","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"osworld-v2@2026.06.24","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5cf3e8e55a6bd9a8a68a","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.anthropic.com/news/claude-opus-4-5","harness":"Claude Opus 4.5","harnessId":"Claude Opus 4.5","httpStatus":200,"id":"pub-1c1779cd9192c641b96765f2","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Claude Opus 4.5","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"66.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0b2584eb5eaf50661a419ace9b4288786a9205f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3997,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"os_world_external.csv:row=58;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3997,"metadata":{"epoch_file":"os_world_external.csv","notes":"","organization":"Anthropic","release_date":"2025-11-24","source_link":"https://www.anthropic.com/news/claude-opus-4-5"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":66.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"osworld","benchmarkName":"OSWorld 2.0","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"osworld-v2@2026.06.24","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2ffa35e51c9871f9e369","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://os-world.github.io/","harness":"claude-sonnet-4-5-20250929 (15 steps)","harnessId":"claude-sonnet-4-5-20250929 (15 steps)","httpStatus":200,"id":"pub-b785ba471a5adfece1d1250a","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-sonnet-4-5-20250929 (15 steps)","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"42.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e72a96cc4ddaffe616914117690aa938290538e4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3983,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"os_world_external.csv:row=12;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3983,"metadata":{"epoch_file":"os_world_external.csv","notes":"","organization":"Anthropic","release_date":"2025-09-29","source_link":"https://os-world.github.io/"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":42.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"osworld","benchmarkName":"OSWorld 2.0","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"osworld-v2@2026.06.24","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e18ea945d88b5c6cd05b","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://os-world.github.io/","harness":"claude-sonnet-4-5-20250929 (50 steps)","harnessId":"claude-sonnet-4-5-20250929 (50 steps)","httpStatus":200,"id":"pub-b0bbb6213344156f693fbf72","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-sonnet-4-5-20250929 (50 steps)","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"58.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e72a96cc4ddaffe616914117690aa938290538e4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3981,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"os_world_external.csv:row=6;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3981,"metadata":{"epoch_file":"os_world_external.csv","notes":"","organization":"Anthropic","release_date":"2025-09-29","source_link":"https://os-world.github.io/"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":58.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"osworld","benchmarkName":"OSWorld 2.0","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"osworld-v2@2026.06.24","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-094a3263cfa0fa0caa33","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://os-world.github.io/","harness":"claude-sonnet-4-5-20250929 (100 steps)","harnessId":"claude-sonnet-4-5-20250929 (100 steps)","httpStatus":200,"id":"pub-a77e5ebf1c817e7c20ec3137","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-sonnet-4-5-20250929 (100 steps)","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"62.9","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e72a96cc4ddaffe616914117690aa938290538e4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3980,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"os_world_external.csv:row=5;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3980,"metadata":{"epoch_file":"os_world_external.csv","notes":"","organization":"Anthropic","release_date":"2025-09-29","source_link":"https://os-world.github.io/"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":62.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"osworld","benchmarkName":"OSWorld 2.0","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"osworld-v2@2026.06.24","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b36ed1d0c003bcfe090c","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://os-world.github.io/","harness":"claude-sonnet-4-6 (100 steps)","harnessId":"claude-sonnet-4-6 (100 steps)","httpStatus":200,"id":"pub-d14d14cff8dcec07907876bc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"claude-sonnet-4-6 (100 steps)","subject_type":"system"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"72.1","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"014dc060e2ea77e718dc140e9cc7c7a4aa933570","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3979,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"os_world_external.csv:row=3;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3979,"metadata":{"epoch_file":"os_world_external.csv","notes":"","organization":"Anthropic","release_date":"2026-02-17","source_link":"https://os-world.github.io/"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":72.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"osworld","benchmarkName":"OSWorld 2.0","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"osworld-v2@2026.06.24","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-311857bad49e223f4552","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://os-world.github.io/","harness":"Kimi-K2.5","harnessId":"Kimi-K2.5","httpStatus":200,"id":"pub-6f83646fcb0f5ed105b33efa","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"Score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":"Kimi-K2.5","subject_type":"system"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unit_inferred_from_column"],"rank":null,"rawValue":"63.3","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4d8aad5bfb1e2422d06621cb4aefdd7d5ca384ab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3998,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"os_world_external.csv:row=59;column=Score","sourceMetricId":"Score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":3998,"metadata":{"epoch_file":"os_world_external.csv","notes":"","organization":"Moonshot","release_date":"2026-01-27","source_link":"https://os-world.github.io/"},"sourceStatus":"candidate"},"sourceSubjectType":"system","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":63.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-05-28","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-91b0ceced0658be757ad","canonicalModelId":"deepseek/r1@0528","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-736b6b0f079dbfb6bebc50be","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek-R1-0528","modelRef":"DeepSeek-R1-0528","observedAt":"2025-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.274","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fb47a31069ae5f728fdb53bb8cf89b13f50f9607","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4767,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=78;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4767,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"DeepSeek","release_date":"2025-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":27.400000000000002,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-06-09","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-78ec4510dfe85465a3f9","canonicalModelId":"anthropic/claude-fable-5@2026-06-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0ff91d42dd0aacc4e98b3f34","mappingCandidates":["anthropic/claude-fable-5@2026-06-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Fable 5","modelRef":"claude-fable-5_xhigh","observedAt":"2026-06-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.683","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58609a25c593c2c6916a20312a8d3fa5a69c660f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4717,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4717,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.30000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-10-15","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dc39f408224e17f537f9","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3722ac0dcbfccae4a48645bc","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Haiku 4.5","modelRef":"claude-haiku-4-5-20251001_32K","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.059","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc5f3b6cd805d47460f51ff7e2e8ec498d36e03f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4757,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=68;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4757,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-10-15","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":5.8999999999999995,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b7a7407df95890589e37","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e28065e8e5a75e43b1b12e39","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101_32K","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.418","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"593ebc52dbcbe8b905d08b98c252795d635f4fc0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4759,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=70;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4759,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8844463995c756d8c4b8","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-c2b24b0ec9274d4b42ae1196","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.431","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f1a737ece3c2c8389cf8a28eeccfb825ed30ae0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4740,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=51;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4740,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0824b8a6ece6c3b448b8","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a057096f9c8b853ebfeb093f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_32K","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.464859437751004","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f1a737ece3c2c8389cf8a28eeccfb825ed30ae0","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4739,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=50;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4739,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.4859437751004,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-775d15c0ccdd2a32a845","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-74cd2107e82eb2e3ceaaef64","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6_max","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.41041041041041043","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2f1a737ece3c2c8389cf8a28eeccfb825ed30ae0","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4738,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=49;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4738,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.04104104104104,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9cab8a92e416a70236a8","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-397fe6a492a3fcd8a453d0f6","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_xhigh","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.506","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4cad22ac40f9d0de98e167fe47d42cddc7d7d9b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4729,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=40;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4729,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-05-28","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b8ed9d2ab26bf4d7e7f8","canonicalModelId":"anthropic/claude-opus-4.8@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-383d1d884c4ca7f49c0ba970","mappingCandidates":["anthropic/claude-opus-4.8@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.8","modelRef":"claude-opus-4-8_max","observedAt":"2026-05-28","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.395","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b6e44a43c281c3b5d800aff60962073b6764587a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4718,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=29;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4718,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-05-28","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-24","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ba7401fc861118eba0c9","canonicalModelId":"anthropic/claude-opus-5@2026-07-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9037b7b1b67525a4190f0b19","mappingCandidates":["anthropic/claude-opus-5@2026-07-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 5","modelRef":"claude-opus-5_max","observedAt":"2026-07-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.567","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a6bc5d76d0304ee49a61a2e8c92d35e509525d82","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4702,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=13;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4702,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-07-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-473ac538e85da9f4452e","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a0cde89a800174adabce94eb","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.13","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8190f7889c7acbd4b02e6fd083789ffaff0c687","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4765,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=76;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4765,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":13.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5f14e4f352842fafe435","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-aba07499ffe4ed851b8470ac","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929_59K","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.236","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c8190f7889c7acbd4b02e6fd083789ffaff0c687","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4753,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=64;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4753,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":23.599999999999998,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7d3988a0fa9da8caaca0","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-77e4c54b2e18893457dcf8f3","mappingCandidates":["anthropic/claude-sonnet-4.6@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6_32K","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.29","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"207204564ca219215c22943136b568a4a924ba6a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4736,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=47;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4736,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.999999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-06-30","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a8ba382f25a6b7c1b3e9","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6cadcc6de9dc2d361f7bfb4c","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_max","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.223","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1c51449d7fc947ea890594d4227ba0cf319f92","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4711,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=22;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4711,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":22.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-06-30","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b5f4b2405be0671aba5f","canonicalModelId":"anthropic/claude-sonnet-5@2026-06-30","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ff8a1d9894b509dd010bf49","mappingCandidates":["anthropic/claude-sonnet-5@2026-06-30"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 5","modelRef":"claude-sonnet-5_xhigh","observedAt":"2026-06-30","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.25","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dd1c51449d7fc947ea890594d4227ba0cf319f92","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4710,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=21;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4710,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-06-30","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":25.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-31","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-56f297c9c19230066308","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ed583a5959fd7e45cd2a8a6f","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-v4-flash-0731_max","observedAt":"2026-07-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.3466933867735471","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cb7527ec2e6a63b39362d1d9569eeb4784b1cb51","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4700,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4700,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"DeepSeek","release_date":"2026-07-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.66933867735471,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-1e0e10fbd9181595b4eb","canonicalModelId":"deepseek/v4-pro@2026-08-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-7c33198fdc14fb899e11f636","mappingCandidates":["deepseek/v4-pro@2026-08-13"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro (0813)","modelRef":"deepseek-v4-pro-0813_max","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5546639919759278","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0caf931ddb3dbf4a730dc4d4c02b086fe192a18a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4692,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4692,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"DeepSeek","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":55.46639919759278,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-04-24","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-664742aa1d3ed6eaee7e","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8d8814e6d20f043f8e6424e0","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.57","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"83cf72a0a7552c6f1eea30deec0b66fd34e85738","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4714,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=25;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4714,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.99999999999999,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-b453573106d63c3793aa","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4256d134d49258f051dae7fc","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Epoch provider-qualified spelling of Kimi K2.5.","registry":"model_aliases.json"},"mappingNote":"Epoch provider-qualified spelling of Kimi K2.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"fireworks/kimi-k2p5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.339","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9817c1407e53e8916a1939f1a581fa2ffb5269b6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4742,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=53;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4742,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":33.900000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-981ff71aff7c1362dcd0","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2c6bcb4c6a6a45d4f41de93a","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.56","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f039a23b494c755a23eafe068183bd6d1a658937","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4766,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=77;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4766,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.00000000000001,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-19","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2f312db458b1ae051d34","canonicalModelId":"google/gemini-3.1-pro@2026-02-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec350ee762bca590808b305f","mappingCandidates":["google/gemini-3.1-pro@2026-02-19"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.1 Pro","modelRef":"gemini-3.1-pro-preview","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.773","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a3e0cb082048e2c9da82531c1bcaf35a73130f0c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4737,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=48;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4737,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-02-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f922ab7d1f7abfec4549","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-65acca8e768885b2f9617882","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.684","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbbb1c6a7f375f368e560ce8f545320b49f1724d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4719,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4719,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-21","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a7ce2829c46e26465012","canonicalModelId":"google/gemini-3.6-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-20f139b3f51109487fde834d","mappingCandidates":["google/gemini-3.6-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.6 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.6 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.6 Flash","modelRef":"gemini-3.6-flash_high","observedAt":"2026-07-21","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.687","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ac63b3020dfb49eefcdec4b913659990896a1a69","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4701,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=12;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4701,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-07-21","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":68.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-08-13","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-da97e7146c9140b15e87","canonicalModelId":"google/gemini-3.7-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6ad57ffc6b0d5fcc9101af29","mappingCandidates":["google/gemini-3.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.7 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.7 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.7 Flash","modelRef":"gemini-3.7-flash_high","observedAt":"2026-08-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.712","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0d02c5798148ebd29aa28b0463766d8de4d9c7a1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4693,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4693,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Google DeepMind","release_date":"2026-08-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-12-22","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-e9aeb284a445eaea06bf","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1b314af0cdab4ca8744677e4","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-4.7","modelRef":"glm-4.7","observedAt":"2025-12-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.315","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"79fd975e36c8ace6dcc563abb1e6e333f30a7ee0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4741,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=52;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4741,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2025-12-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":31.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-04-07","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ed30ab59278abccbc331","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6c48e6456800b201a519629f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.37286432160804023","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"503db941b11116a5cbda750f15d20e126582b3be","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4725,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4725,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":37.286432160804026,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-06-16","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-eb1d5a5ed73e23865bec","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-18e646005d7f34be1170d703","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.381","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c42f203b5f25879120b883572adf30037bb52870","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4712,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=23;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4712,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-08-14","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2f2650326f4b5e2b6bab","canonicalModelId":"zai/glm-5.3@2026-08-14","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3cadfa09656efff3431721e9","mappingCandidates":["zai/glm-5.3@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.3","modelRef":"glm-5.3_max","observedAt":"2026-08-14","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.402","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a35fc4ec0e79283704a8213a22dbc0cacd3c85db","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4691,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4691,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-08-14","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":40.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5e8e77b04afe423bed21","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-768372220e23479fd3c30009","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.506","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"09d5a9c87fbb1c5861383c02795bde505790f661","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4763,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=74;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4763,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":50.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-acbc8f5866a4eb26f073","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-83b7606527fe0dc28768dfd1","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.21","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2d4c4ab4609153e62434707c0a41f9e6f4a95b10","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4756,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=67;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4756,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d9cd1d8d53f1ca777f36","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-27b523aaa2b652ab1cb2dce7","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 nano release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 nano release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 nano","modelRef":"gpt-5-nano-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.122","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5b60a2b6871c69e515ca74ec5ee5eedf537ccb9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4761,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=72;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4761,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6a3647a0f8d45f66815f","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-294f1a876fe9d4bf38c2ef8d","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.489","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a4c9b0dba11438464eaa5c8ae7ac66f6a99ccdb7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4754,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=65;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4754,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":48.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-33ca1b34e4b6aa42e281","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9bf2b4d75cf85cb78a54ac04","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.382","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c10217802faf48d99f6f08b9213bded7a0e564d9","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4748,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=59;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4748,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-85ddb52360f24606471e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f356348a4f0e622514f3798c","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_low","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.347","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c10217802faf48d99f6f08b9213bded7a0e564d9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4750,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=61;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4750,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":34.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-16a357599a102c9ca59f","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-9131b6b86b7ace0b76000a63","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_medium","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.354","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c10217802faf48d99f6f08b9213bded7a0e564d9","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4747,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=58;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4747,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":35.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f39a8c614bb568f6c430","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2eb8d346a9eb2b8fc5c836a0","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_xhigh","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.389","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c10217802faf48d99f6f08b9213bded7a0e564d9","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4749,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=60;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4749,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-775f613a0068a908eb32","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-78db8af49712adde577504f0","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4483450351053159","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14fbbceca61ba55cd193f7fb67d4aa5d084b64fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4735,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=46;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4735,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":44.83450351053159,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-65d7ac96fd70fb34e4ed","canonicalModelId":"openai/gpt-5.4-mini@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-60e68f41fed5be18f7dbc649","mappingCandidates":["openai/gpt-5.4-mini@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-17 GPT-5.4 mini release with an effort suffix.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 mini","modelRef":"gpt-5.4-mini-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.286","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"71ccb69a5974258ca61310c0d4f2b5f075d08594","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4730,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=41;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4730,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":28.599999999999998,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-03-17","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-547d6549090273b57b55","canonicalModelId":"openai/gpt-5.4-nano@2026-03-17","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d8162f874804661ff1ba156c","mappingCandidates":["openai/gpt-5.4-nano@2026-03-17"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the explicit 2026-03-17 GPT-5.4 nano snapshot.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 nano","modelRef":"gpt-5.4-nano-2026-03-17_high","observedAt":"2026-03-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.12","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d2bd8fcba98fe3ac0e10289a01bdde719901d9a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4731,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=42;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4731,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":12.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-61009e344141705ad0e7","canonicalModelId":"openai/gpt-5.4-pro@2026-03-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e2eb45294f8271fdef151f5d","mappingCandidates":["openai/gpt-5.4-pro@2026-03-05"],"mappingEvidence":{"confidence":"high","note":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch spellings for the dated GPT-5.4 Pro release; compute and web-app settings remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4 Pro","modelRef":"gpt-5.4-pro-2026-03-05_xhigh","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.478","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f34cbf8dba3c525f59bfa721c40dbd8e84cea775","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4734,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=45;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4734,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":47.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-05-05","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5aaaf09e6ae932d05ff8","canonicalModelId":"openai/gpt-5.5@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2870bdd618a8dbccbe964988","mappingCandidates":["openai/gpt-5.5@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.5","modelRef":"gpt-5.5-instant","observedAt":"2026-05-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.38","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e5fd62b71a2522f59fac875fc54116fb61939795","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4699,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=10;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4699,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-05-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9cda8f107e750adc4e27","canonicalModelId":"openai/gpt-5.6-luna@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1dae7a5bd43139af9bf9fc6a","mappingCandidates":["openai/gpt-5.6-luna@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Luna","modelRef":"gpt-5.6-luna_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"769b2fa0458e14da0a477d9794daeec61cf75a3f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4707,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4707,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":41.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-8f57d99f71fa3232c506","canonicalModelId":"openai/gpt-5.6-sol@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b659f488fbe8c09144a121be","mappingCandidates":["openai/gpt-5.6-sol@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Sol","modelRef":"gpt-5.6-sol_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.716","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dbf326f07fa938ec13f64a79dda9dde5e923e904","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4708,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4708,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-09","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-dca25925333011e75cc7","canonicalModelId":"openai/gpt-5.6-terra@2026-07-09","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1dbed4d1386f82ed68e48818","mappingCandidates":["openai/gpt-5.6-terra@2026-07-09"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.6 Terra","modelRef":"gpt-5.6-terra_max","observedAt":"2026-07-09","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.431","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3e29f0818cf9c311ad34db007eb0828d4f54374b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4706,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4706,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-07-09","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":43.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-08","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-00133bac2829fa81d6b4","canonicalModelId":"xai/grok-4.5@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-428afedf5a5da7fa50c1ce26","mappingCandidates":["xai/grok-4.5@2026-07-16"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.5.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.5","modelRef":"grok-4.5_high","observedAt":"2026-07-08","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.535","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"731882f693a7b6cfa1afcd862f4cf3898f0bfbbd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4709,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4709,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"xAI","release_date":"2026-07-08","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-97f8c99592e48643ac4c","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-3639210223e77260d3afd8e6","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_high","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.53","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5746c9aa5fe1d6498b83b8c2b73c7b790e29c90","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4696,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4696,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":53.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-08-12","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-68a5250d0370084dda3e","canonicalModelId":"xai/grok-4.6@2026-08-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-52e58a93ef9e1f4243eee16b","mappingCandidates":["xai/grok-4.6@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Grok 4.6.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Grok 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Grok 4.6","modelRef":"grok-4.6_xhigh","observedAt":"2026-08-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.544","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b5746c9aa5fe1d6498b83b8c2b73c7b790e29c90","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4695,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4695,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"xAI","release_date":"2026-08-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":54.400000000000006,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-136011fae14fabb85823","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a83f0007f2f0912b1d52871d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.387","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1ac31824b0e78702084e1a5f2709dd5ff010459e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4726,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=37;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4726,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":38.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-06-12","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3e798af960915b0bf570","canonicalModelId":"moonshot/kimi-k2.7-code@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4220fcaf105855651be90c5b","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.7 Code","modelRef":"kimi-k2.7-code","observedAt":"2026-06-12","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.392","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9ffc58d0af5092ec36a643009a8fc33db7ef2a1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4716,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=27;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4716,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Moonshot","release_date":"2026-06-12","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":39.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-07-16","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-78783c6fe27e41eb5e17","canonicalModelId":"moonshot/kimi-k3@2026-07-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-16d06a53fb538c68bf6ef8d2","mappingCandidates":["moonshot/kimi-k3@2026-07-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K3","modelRef":"kimi-k3_max","observedAt":"2026-07-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.427","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f3eb6b8564a2cd6d285b7eb46b56b0ef3ac98150","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4703,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=14;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4703,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Moonshot","release_date":"2026-07-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":42.699999999999996,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-25","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0b3318f998b410e4deaa","canonicalModelId":"qwen/qwen3.5-flash@2026-02-25","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ec54ec0c99c9fde5edef4769","mappingCandidates":["qwen/qwen3.5-flash@2026-02-25"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Flash","modelRef":"qwen3.5-flash","observedAt":"2026-02-25","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.19759277833500502","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cff8e7c464581ab038d150ed79cb346d35d73a11","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4724,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=35;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4724,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-25","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":19.7592778335005,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-02-16","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3b37d990211e4449f3f1","canonicalModelId":"qwen/qwen3.5-plus@2026-02-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0f9600e4462c6cbb651a8f1b","mappingCandidates":["qwen/qwen3.5-plus@2026-02-16"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.5 Plus","modelRef":"qwen3.5-plus","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.2600401606425703","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"21baec2df927a7202f9a0bf935b9e54846b63c79","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4721,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4721,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-02-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":26.00401606425703,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-04-27","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-ca91acd6ef389ca52f04","canonicalModelId":"qwen/qwen3.6-flash@2026-04-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2ac59fb8697af9dbba7c791f","mappingCandidates":["qwen/qwen3.6-flash@2026-04-27"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Flash","modelRef":"qwen3.6-flash","observedAt":"2026-04-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.21163490471414242","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22e9580cfbc9fb2770c283612ab0dcf928aee88a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4723,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4723,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":21.163490471414242,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9bb4cea18568815da36b","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-711c24e08dc351d6b4fd7143","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5692771084337349","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"22cc30fd2afd19ddd4886c65d7a0d26fee9fb0a8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4720,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4720,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":56.92771084337349,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-03-31","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-abd22156a39b3de0db0e","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-fe92a415d9b118e1ec1083e9","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.4909819639278557","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fc267bc88b94963b990e80975f7ddd82f92cecac","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4722,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4722,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":49.09819639278557,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-d1677c9541ef11ccbfa6","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-67fb059c049115db34050519","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5851703406813628","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cc851d71910e80a1afa8a4a9554ee20ab0da745d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4715,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=26;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4715,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":58.517034068136276,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"simpleqa","benchmarkName":"SimpleQA","benchmarkVersion":"2026-08-02","benchmarkVersionHint":"simpleqa@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4b0929de6e48390345ca","canonicalModelId":"qwen/qwen3.8-max@2026-08-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5ba3e038716f83bce20485c4","mappingCandidates":["qwen/qwen3.8-max@2026-08-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.8 Max","modelRef":"qwen3.8-max_xhigh","observedAt":"2026-08-02","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.46292585170340683","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"31c55b704d2497c153573ddaefbfa59a119dcc2c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4698,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"simpleqa_verified.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4698,"metadata":{"epoch_file":"simpleqa_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-08-02","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":46.292585170340686,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-fb4ac241a2733a82af12","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-c23eb2217cb9f8cf1d7f71e5","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus","observedAt":"2025-08-02","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias"],"rank":17,"rawValue":67.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c85ebb6e85f92b3f279b527e03ca6a9652439b43","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":17,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[16]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":17,"metadata":{"agent":"mini-SWE-agent","cost":565.6352234999999,"display_name":"Claude 4 Opus (20250514)","folder":"20250802_mini-v1.0.0_claude-4-opus-20250514","instance_calls":31.076,"logs":"s3://swe-bench-submissions/bash-only/20250802_mini-v1.0.0_claude-4-opus-20250514/logs","model_display":"Claude 4 Opus","model_org":"Anthropic","model_release_date":20250522,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250802_mini-v1.0.0_claude-4-opus-20250514/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":67.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-7951b632bf52d620ec89","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3b2c584f5ebd8c8caeb2c55a","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Haiku 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Haiku 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":18,"rawValue":66.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d1f79d5c7b0c527093211c147306b74dfe9ea5a3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":18,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[17]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":18,"metadata":{"agent":"mini-SWE-agent","cost":165.46221805000002,"display_name":"Claude 4.5 Haiku (high)","folder":"20260217_mini-v2.0.0_claude-4-5-haiku-high","instance_calls":66.154,"logs":null,"model_display":"Claude 4.5 Haiku","model_org":"Anthropic","model_release_date":20251015,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_claude-4-5-haiku-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-0f620c0e2bb566473098","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3348e47773699ff4c046b6ca","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":1,"rawValue":76.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d053db7db579b9dc834bfb2bdb68807fb8b1a245","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":1,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[0]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":1,"metadata":{"agent":"mini-SWE-agent","cost":376.9539985,"display_name":"Claude 4.5 Opus (high)","folder":"20260217_mini-v2.0.0_claude-4-5-opus-high","instance_calls":32.896,"logs":null,"model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":20251124,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_claude-4-5-opus-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":76.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-e9e07391a91eed1a7a8f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-885195423c2d9360149f63a8","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias"],"rank":5,"rawValue":74.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d053db7db579b9dc834bfb2bdb68807fb8b1a245","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":5,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[4]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":5,"metadata":{"agent":"mini-SWE-agent","cost":360.6224985,"display_name":"Claude 4.5 Opus (20251101) (medium)","folder":"20251124_mini-v1.16.0_claude-opus-4-5-20251101","instance_calls":38.042,"logs":"s3://swe-bench-submissions/bash-only/20251124_mini-v1.16.0_claude-opus-4-5-20251101/logs","model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":20251124,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20251124_mini-v1.16.0_claude-opus-4-5-20251101/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":74.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-817b9792f551b901c1a3","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-4c214e724420dcd6354e21e7","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Sonnet 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Sonnet 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":11,"rawValue":71.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cfd1298529eb33987fe90a974fb07f59ea8a77da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":11,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[10]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":11,"metadata":{"agent":"mini-SWE-agent","cost":328.9482492,"display_name":"Claude 4.5 Sonnet (high)","folder":"20260217_mini-v2.0.0_claude-4-5-sonnet-high","instance_calls":48.3,"logs":null,"model_display":"Claude 4.5 Sonnet","model_org":"Anthropic","model_release_date":20250929,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_claude-4-5-sonnet-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":71.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-dbb9bc6865b3b2c50ece","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-5bbc2f607972f47610fad570","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Sonnet 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Sonnet 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias"],"rank":13,"rawValue":70.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cfd1298529eb33987fe90a974fb07f59ea8a77da","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":13,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[12]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":13,"metadata":{"agent":"mini-SWE-agent","cost":279.16737045,"display_name":"Claude 4.5 Sonnet (20250929)","folder":"20250929_mini-v1.13.3_sonnet-4-5-20250929","instance_calls":50.988,"logs":"s3://swe-bench-submissions/bash-only/20250929_mini-v1.13.3_sonnet-4-5-20250929/logs","model_display":"Claude 4.5 Sonnet","model_org":"Anthropic","model_release_date":20250929,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250929_mini-v1.13.3_sonnet-4-5-20250929/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-28518abbc5fad6098a67","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-653ea4df2110cb09ad8431de","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.6.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.6","modelRef":"Claude 4.6 Opus","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":4,"rawValue":75.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"93f91adf979d530947d7006b53d95fce5c6f9e6d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":4,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[3]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":4,"metadata":{"agent":"mini-SWE-agent","cost":275.76132225,"display_name":"Claude 4.6 Opus","folder":"20260217_mini-v2.0.0_claude-4-6-opus","instance_calls":28.932,"logs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_claude-4-6-opus/logs","model_display":"Claude 4.6 Opus","model_org":"Anthropic","model_release_date":20260205,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_claude-4-6-opus/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":75.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-6d7f58b8053cbe938ca5","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-7eca0998b99d66e40564340f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"DeepSeek V3.2","modelRef":"DeepSeek V3.2","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["not_independently_checked"],"rank":14,"rawValue":70.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3140f74bd9fd8a3f6da0d138ae9e9f60a6a0953c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":14,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[13]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":14,"metadata":{"agent":"mini-SWE-agent","cost":223.92324825,"display_name":"DeepSeek V3.2 (high)","folder":"20260217_mini-v2.0.0_deepseek-3-2-high","instance_calls":88.506,"logs":null,"model_display":"DeepSeek V3.2","model_org":"DeepSeek","model_release_date":20251201,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_deepseek-3-2-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-0399279f7e011b4effc8","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-270e39ae065a85e897f8505c","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of the DeepSeek V3.2 reasoning endpoint; the harness stays on the source row.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of the DeepSeek V3.2 reasoning endpoint; the harness stays on the source row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"DeepSeek V3.2","modelRef":"DeepSeek V3.2 Reasoner","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias"],"rank":25,"rawValue":60.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3140f74bd9fd8a3f6da0d138ae9e9f60a6a0953c","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":25,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[24]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":25,"metadata":{"agent":"mini-SWE-agent","cost":14.0381514,"display_name":"DeepSeek V3.2 Reasoner","folder":"20251201_mini-v1.17.1_deepseek-v3.2-reasoner","instance_calls":46.404,"logs":"s3://swe-bench-submissions/bash-only/20251201_mini-v1.17.1_deepseek-v3.2-reasoner/logs","model_display":"DeepSeek V3.2 Reasoner","model_org":"deepseek","model_release_date":20251201,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20251201_mini-v1.17.1_deepseek-v3.2-reasoner/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-3c5eb6eb19a151e3d07f","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-8eff87b3391725b0b2cc81be","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GLM-5","modelRef":"GLM 5","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","not_independently_checked"],"rank":8,"rawValue":72.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f75a4cc6e4e6e82d10a75d346c63f556ed6a3d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":8,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[7]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":8,"metadata":{"agent":"mini-SWE-agent","cost":267.19456811,"display_name":"GLM 5 (high)","folder":"20260217_mini-v2.0.0_glm-5-high","instance_calls":76.176,"logs":null,"model_display":"GLM 5","model_org":"Z-AI","model_release_date":20260211,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_glm-5-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-03f18c03a65007abb456","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-7c2a25c5f63652a82ac28dbe","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5","modelRef":"GPT 5","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping"],"rank":21,"rawValue":65.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa4b77bc7665f6bc658244089d3719d20c876c52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":21,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[20]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":21,"metadata":{"agent":"mini-SWE-agent","cost":140.19150875,"display_name":"GPT 5 (2025-08-07) (medium)","folder":"20250807_mini-v1.7.0_gpt-5","instance_calls":13.208,"logs":"s3://swe-bench-submissions/bash-only/20250807_mini-v1.7.0_gpt-5/logs","model_display":"GPT 5","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250807_mini-v1.7.0_gpt-5/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-2bd824535f050d2e76e2","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-4f629a205fe1df8a0fa6d072","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 mini","modelRef":"GPT 5 mini","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping"],"rank":26,"rawValue":59.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b695b769ea10e457a54972b7a7f78f0f2df726b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":26,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[25]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":26,"metadata":{"agent":"mini-SWE-agent","cost":17.73853365,"display_name":"GPT 5 mini (2025-08-07) (medium)","folder":"20250807_mini-v1.7.0_gpt-5-mini","instance_calls":14.466,"logs":"s3://swe-bench-submissions/bash-only/20250807_mini-v1.7.0_gpt-5-mini/logs","model_display":"GPT 5 mini","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250807_mini-v1.7.0_gpt-5-mini/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":59.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-8211c45ac84cff5ad214","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-87992e477f49653ed9a8574b","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 mini","modelRef":"GPT 5 mini","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping","not_independently_checked"],"rank":29,"rawValue":56.2,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2b695b769ea10e457a54972b7a7f78f0f2df726b","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":29,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[28]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":29,"metadata":{"agent":"mini-SWE-agent","cost":23.600609549999998,"display_name":"GPT 5 mini","folder":"20260217_mini-v2.0.0_gpt-5-mini","instance_calls":20.342,"logs":null,"model_display":"GPT 5 mini","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_gpt-5-mini/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":56.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-cd673e58f9a95a2fe6a2","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-ea871f28f6d39c9013120f83","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 nano","modelRef":"GPT 5 nano","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping"],"rank":39,"rawValue":34.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4c928f678bcc86d49bac38c530dcb5ef1005da67","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":39,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[38]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":39,"metadata":{"agent":"mini-SWE-agent","cost":19.037590390000002,"display_name":"GPT 5 nano (2025-08-07) (medium)","folder":"20250807_mini-v1.7.0_gpt-5-nano","instance_calls":39.76,"logs":"s3://swe-bench-submissions/bash-only/20250807_mini-v1.7.0_gpt-5-nano/logs","model_display":"GPT 5 nano","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250807_mini-v1.7.0_gpt-5-nano/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":34.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-8634b955f73469ea5cb2","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-6a92868a9aa5dbf33c16d52e","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.1","modelRef":"GPT 5.1","observedAt":"2025-11-20","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":20,"rawValue":66.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"402ecb6467ac35b18e04cb726912a652255f4ffd","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":20,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[19]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":20,"metadata":{"agent":"mini-SWE-agent","cost":153.07773625000002,"display_name":"GPT 5.1 (2025-11-13) (medium)","folder":"20251120_mini-v1.15.0_gpt-5.1-2025-11-13","instance_calls":21.0,"logs":"s3://swe-bench-submissions/bash-only/20251120_mini-v1.15.0_gpt-5.1-2025-11-13/logs","model_display":"GPT 5.1","model_org":"OpenAI","model_release_date":20251112,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20251120_mini-v1.15.0_gpt-5.1-2025-11-13/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-8d207fe2543eb89fb0fc","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-ce878bc6e4f77aa8d6ad30d8","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.1 Codex","modelRef":"GPT 5.1 Codex","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":19,"rawValue":66.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"35ed5b17c2c4ea32cc2e31da952b79d81a05a6ed","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":19,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[18]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":19,"metadata":{"agent":"mini-SWE-agent","cost":294.44117500000004,"display_name":"GPT 5.1 Codex (medium)","folder":"20251124_mini-v1.16.0_gpt-5.1-codex","instance_calls":23.882,"logs":"s3://swe-bench-submissions/bash-only/20251124_mini-v1.16.0_gpt-5.1-codex/logs","model_display":"GPT 5.1 Codex","model_org":"OpenAI","model_release_date":20251112,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20251124_mini-v1.16.0_gpt-5.1-codex/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-9f1421b75723ac431c61","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-ba9fd9c49dad4196ac2b055d","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":9,"rawValue":72.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4dfc0ed242c3db2cc176e8529689597420648f1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":9,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[8]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":9,"metadata":{"agent":"mini-SWE-agent","cost":236.78350885,"display_name":"GPT 5.2 (high)","folder":"20260217_mini-v2.0.0_gpt-5-2-high","instance_calls":35.046,"logs":null,"model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_gpt-5-2-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-2d18926eac28ce055ed0","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-7f7b84e82d507a3dc6508fce","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":10,"rawValue":71.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4dfc0ed242c3db2cc176e8529689597420648f1","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":10,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[9]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":10,"metadata":{"agent":"mini-SWE-agent","cost":260.1295018,"display_name":"GPT 5.2 (2025-12-11) (high)","folder":"20251211_mini-v1.17.2_gpt-5.2-2025-12-11-high","instance_calls":19.762,"logs":"s3://swe-bench-submissions/bash-only/20251211_mini-v1.17.2_gpt-5.2-2025-12-11-high/logs","model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20251211_mini-v1.17.2_gpt-5.2-2025-12-11-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":71.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-d771fc2131460a8c42a8","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-b191347af15a05285e9d15c6","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":16,"rawValue":69.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b4dfc0ed242c3db2cc176e8529689597420648f1","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":16,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[15]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":16,"metadata":{"agent":"mini-SWE-agent","cost":134.8337487,"display_name":"GPT 5.2 (2025-12-11)","folder":"20251211_mini-v1.17.2_gpt-5.2-2025-12-11","instance_calls":16.27,"logs":"s3://swe-bench-submissions/bash-only/20251211_mini-v1.17.2_gpt-5.2-2025-12-11/logs","model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20251211_mini-v1.17.2_gpt-5.2-2025-12-11/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":69.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-60265243f183186b7e12","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-1d162d32d186598ea0f0a1ee","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2 Codex","modelRef":"GPT 5.2 Codex","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":7,"rawValue":72.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ad68bcc8e7783b273853ef8fab1d0f6db53b69b8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":7,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[6]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":7,"metadata":{"agent":"mini-SWE-agent","cost":224.71194935,"display_name":"GPT 5.2 Codex","folder":"20260219_mini-v2.0.0_gpt-5-2-codex","instance_calls":28.072,"logs":"s3://swe-bench-submissions/bash-only/20260219_mini-v2.0.0_gpt-5-2-codex/logs","model_display":"GPT 5.2 Codex","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260219_mini-v2.0.0_gpt-5-2-codex/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-25bf4920d3b8207b1a1c","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-7266a3ec84cb75685113a084","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Gemini 2.5 Flash","modelRef":"Gemini 2.5 Flash","observedAt":"2025-07-26","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score"],"rank":40,"rawValue":28.73,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"826f23869bb0339ebe773cceb6be18a2985d50fa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":40,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[39]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":40,"metadata":{"agent":"mini-SWE-agent","cost":66.942859175,"display_name":"Gemini 2.5 Flash (2025-04-17)","folder":"20250726_mini-v1.0.0_gemini-2.5-flash","instance_calls":26.646,"logs":"s3://swe-bench-submissions/bash-only/20250726_mini-v1.0.0_gemini-2.5-flash/logs","model_display":"Gemini 2.5 Flash","model_org":"Google DeepMind","model_release_date":20250417,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250726_mini-v1.0.0_gemini-2.5-flash/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":28.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-42a4db817183c55a426c","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-cdba0cc817486f9dea38693d","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Gemini 2.5 Pro","modelRef":"Gemini 2.5 Pro","observedAt":"2025-07-26","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score"],"rank":34,"rawValue":53.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"15346035eb2fb3922d4620e74365925ba36c1759","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":34,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[33]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":34,"metadata":{"agent":"mini-SWE-agent","cost":144.18604781250002,"display_name":"Gemini 2.5 Pro (2025-05-06)","folder":"20250726_mini-v1.0.0_gemini-2.5-pro","instance_calls":20.478,"logs":"s3://swe-bench-submissions/bash-only/20250726_mini-v1.0.0_gemini-2.5-pro/logs","model_display":"Gemini 2.5 Pro","model_org":"Google DeepMind","model_release_date":20250506,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250726_mini-v1.0.0_gemini-2.5-pro/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":53.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-49325b19c2cc48086a94","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-ade8d3f221c939a0dcf0271f","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Kimi K2.5","modelRef":"Kimi K2.5","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","not_independently_checked"],"rank":12,"rawValue":70.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"18383d709ddbf9f40db2a46fc79dee7c88917fbe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":12,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[11]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":12,"metadata":{"agent":"mini-SWE-agent","cost":73.276870372,"display_name":"Kimi K2.5 (high)","folder":"20260217_mini-v2.0.0_kimi-k2-5-high","instance_calls":51.178,"logs":null,"model_display":"Kimi K2.5","model_org":"Moonshot AI","model_release_date":20260127,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_kimi-k2-5-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-9759864b0dea5ccdc17d","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-bb62e052b97cb4aa93db0565","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"MiniMax M2.5","modelRef":"MiniMax M2.5","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["not_independently_checked"],"rank":3,"rawValue":75.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6cb44cfe54502f2ff486df8bdc501887066acb7e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":3,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[2]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":3,"metadata":{"agent":"mini-SWE-agent","cost":36.644993922,"display_name":"MiniMax M2.5 (high)","folder":"20260217_mini-v2.0.0_minimax-2-5-high","instance_calls":60.45,"logs":null,"model_display":"MiniMax M2.5","model_org":"Minimax","model_release_date":20260212,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20260217_mini-v2.0.0_minimax-2-5-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":75.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-bash-only","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-7efd13520418509ae758","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-6b3438cd0566b5c8bad9e2e7","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentations of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentations of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder 480B/A35B Instruct","observedAt":"2025-08-02","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"bash-only","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias"],"rank":30,"rawValue":55.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"aa2d996dbc58124740fa4c0c62941ba0d393f4a5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":30,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[0]=bash-only;results[29]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":30,"metadata":{"agent":"mini-SWE-agent","cost":123.963554,"display_name":"Qwen3-Coder 480B/A35B Instruct","folder":"20250802_mini-v1.0.0_qwen3-coder-480b-a35b-instruct","instance_calls":64.792,"logs":"s3://swe-bench-submissions/bash-only/20250802_mini-v1.0.0_Qwen3-Coder-480B-A35B-Instruct/logs","model_display":"Qwen3-Coder 480B/A35B Instruct","model_org":"Qwen","model_release_date":20250722,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/bash-only/20250802_mini-v1.0.0_Qwen3-Coder-480B-A35B-Instruct/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":55.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-7f7b1b52e7af8683d547","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-5e4368d4334eeb28539b82c0","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Haiku 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Haiku 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":11,"rawValue":64.7,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1a01c2b0e518c03847fa919f3182a81fdf6f705c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":58,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[10]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":58,"metadata":{"agent":"mini-SWE-agent","cost":113.8470847,"display_name":"Claude 4.5 Haiku","folder":"20260213_mini-v2.0.0a0_claude-4-5-haiku","instance_calls":70.14333333333333,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-5-haiku/logs","model_display":"Claude 4.5 Haiku","model_org":"Anthropic","model_release_date":20251015,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-5-haiku/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":64.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-c4aa25a5e9ae7ae85b65","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-5510550ec846e3cf711703b9","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":3,"rawValue":70.7,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed98059756625b2084af72f309ade5397e06546d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":50,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[2]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":50,"metadata":{"agent":"mini-SWE-agent","cost":250.06589275,"display_name":"Claude 4.5 Opus","folder":"20260213_mini-v2.0.0a0_claude-4-5-opus","instance_calls":30.796666666666667,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-5-opus/logs","model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":20251124,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-5-opus/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-79b5f1287a301eba97a6","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3e58721fcbae8e42750fd004","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Sonnet 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Sonnet 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":8,"rawValue":67.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"98f99c59f7d4155d8ff516696f2851075103449c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":55,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[7]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":55,"metadata":{"agent":"mini-SWE-agent","cost":200.8596708,"display_name":"Claude 4.5 Sonnet","folder":"20260213_mini-v2.0.0a0_claude-4-5-sonnet","instance_calls":42.54333333333334,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-5-sonnet/logs","model_display":"Claude 4.5 Sonnet","model_org":"Anthropic","model_release_date":20250929,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-5-sonnet/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":67.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-dc08b5591ba5e3e12ffc","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-b896f746b94c10bbe028c65a","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.6.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.6","modelRef":"Claude 4.6 Opus","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":2,"rawValue":72.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6a004cfe81bea4c10b43f4164a38e51d678f95d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":49,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[1]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":49,"metadata":{"agent":"mini-SWE-agent","cost":198.892073,"display_name":"Claude 4.6 Opus","folder":"20260213_mini-v2.0.0a0_claude-4-6-opus","instance_calls":28.85333333333333,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-6-opus/logs","model_display":"Claude 4.6 Opus","model_org":"Anthropic","model_release_date":20260205,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_claude-4-6-opus/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-3763c28d4f5314b1f5b4","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3d6098563c5debb3ece9c6c0","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"DeepSeek V3.2","modelRef":"DeepSeek V3.2","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":12,"rawValue":59.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"163a56110699d8685eeed6be350398fa2587d6b7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":59,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[11]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":59,"metadata":{"agent":"mini-SWE-agent","cost":115.14765747,"display_name":"DeepSeek V3.2","folder":"20260213_mini-v2.0.0a0_deepseek-3-2","instance_calls":82.65666666666667,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_deepseek-3-2/logs","model_display":"DeepSeek V3.2","model_org":"DeepSeek","model_release_date":20251201,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_deepseek-3-2/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-4a4b0af6523a4a1b514e","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3759d4f70c9feb36ba2413ea","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GLM-5","modelRef":"GLM 5","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":4,"rawValue":69.7,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab7ba6133ae14c55de75e89104b11c63c9009ede","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":51,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[3]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":51,"metadata":{"agent":"mini-SWE-agent","cost":192.74231006,"display_name":"GLM 5","folder":"20260213_mini-v2.0.0a0_glm-5","instance_calls":67.56333333333333,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_glm-5/logs","model_display":"GLM 5","model_org":"Z.ai","model_release_date":20260211,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_glm-5/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":69.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-05aa82d00ebdf8eb22bb","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-beec14f752e63c902e82cb94","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 mini","modelRef":"GPT 5 mini","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":13,"rawValue":39.7,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"044fa2836e22997aea487f3c3d64e62660347969","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":60,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[12]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":60,"metadata":{"agent":"mini-SWE-agent","cost":15.489864099999998,"display_name":"GPT 5 mini","folder":"20260213_mini-v2.0.0a0_gpt-5-mini","instance_calls":30.0,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_gpt-5-mini/logs","model_display":"GPT 5 mini","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_gpt-5-mini/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":39.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-99683e0fa38032b245c5","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-d343190bcd73126efae42d14","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":9,"rawValue":66.7,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"070a84bdf3feff4adac93f1cb431b6ce2112c606","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":56,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[8]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":56,"metadata":{"agent":"mini-SWE-agent","cost":160.92329715,"display_name":"GPT 5.2 (high)","folder":"20260213_mini-v2.0.0a0_gpt-5-2-high","instance_calls":40.266666666666666,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_gpt-5-2-high/logs","model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_gpt-5-2-high/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-1a6a2e71dd71384bd6e8","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-43d23ba49dcce425e7d6f67a","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2 Codex","modelRef":"GPT 5.2 Codex","observedAt":"2026-02-20","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":10,"rawValue":66.3,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"b65a79d6ea691dd4e8a392ea46cfbcfbf6890d77","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":57,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[9]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":57,"metadata":{"agent":"mini-SWE-agent","cost":198.65146735,"display_name":"GPT 5.2 Codex","folder":"20260220_mini-v2.0.0_gpt-5-2-codex","instance_calls":38.17666666666667,"logs":"s3://swe-bench-submissions/multilingual/20260220_mini-v2.0.0_gpt-5-2-codex/logs","model_display":"GPT 5.2 Codex","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260220_mini-v2.0.0_gpt-5-2-codex/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-ade1eae2d2f5e63c4e4f","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-a4152c652f834029fea3e122","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Kimi K2.5","modelRef":"Kimi K2.5","observedAt":"2026-02-13","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":7,"rawValue":67.3,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f7c1e04632ce87b607d2831c14cfaf27f056fde0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":54,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[6]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":54,"metadata":{"agent":"mini-SWE-agent","cost":207.85105504,"display_name":"Kimi K2.5","folder":"20260213_mini-v2.0.0a0_kimi-k2-5","instance_calls":50.43333333333333,"logs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_kimi-k2-5/logs","model_display":"Kimi K2.5","model_org":"Moonshot AI","model_release_date":20260127,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260213_mini-v2.0.0a0_kimi-k2-5/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":67.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-multilingual","benchmarkName":"SWE-bench Multilingual","benchmarkVersion":null,"benchmarkVersionHint":"swebench-multilingual@official","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-21e364d042bf1cdfa01a","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-634fbcab24bf3baced17ad91","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"MiniMax M2.5","modelRef":"MiniMax M2.5","observedAt":"2026-02-16","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Multilingual","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint"],"rank":6,"rawValue":68.3,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"81d8ca91d111f18706705c48e98fc82e7dc3b030","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":53,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[1]=Multilingual;results[5]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":53,"metadata":{"agent":"mini-SWE-agent","cost":30.05951837,"display_name":"MiniMax M2.5","folder":"20260216_mini-v2.0.0a0_minimax-2-5","instance_calls":74.16498316498317,"logs":"s3://swe-bench-submissions/multilingual/20260216_mini-v2.0.0a0_minimax-2-5/logs","model_display":"MiniMax M2.5","model_org":"Minimax","model_release_date":20260212,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/multilingual/20260216_mini-v2.0.0a0_minimax-2-5/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["benchmark:evaluation_mode=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":68.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-d81e87088cbc233c8f06","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://scale.com/leaderboard/swe_bench_pro_public","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-7c3c43c7bd6f107e96855194","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.1","modelRef":"MiniMaxAI/MiniMax-M2.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":38,"rawValue":36.81,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"fadd6592fb490fd1f6924d38c800cbcc979aa37f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":38,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=38;row=37","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":38,"metadata":{"author":"MiniMaxAI","filename":".eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":33,"source_name":"SWE-Bench Pro official evaluation results","source_status":"reported","submission_source_url":"https://scale.com/leaderboard/swe_bench_pro_public"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":36.81,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-8e2f1d78406c67d34caf","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-99fa41299706c19771544a27","mappingCandidates":["minimax/m2.5@2026-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.5","modelRef":"MiniMaxAI/MiniMax-M2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":21,"rawValue":55.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"df4394bd951fcc5ca8bbb96e4ff750cc73bdf7d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":21,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=21;row=20","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":21,"metadata":{"author":"MiniMaxAI","filename":".eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":228703644928,"pull_request":50,"source_name":"Model card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-852ed16516ee82006990","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-612bc9625dcb1142254860a0","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.7","modelRef":"MiniMaxAI/MiniMax-M2.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":18,"rawValue":56.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3effb7bf4ba85c745c3efa5fba3a920df9241079","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":18,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=18;row=17","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":18,"metadata":{"author":"MiniMaxAI","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":13,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-0b5e628b6703972df5d6","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M3","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-983a6cab53e0bb063950db63","mappingCandidates":["minimax/m3@2026-05-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"MiniMaxAI/MiniMax-M3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/minimax-m3.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":10,"rawValue":59,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"09cac6c200c1d5840eedd26c13bc8562b2d59c51","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":10,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=10;row=9","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":10,"metadata":{"author":"MiniMaxAI","filename":".eval_results/minimax-m3.yaml","lower_is_better":false,"num_parameters":427040140160,"pull_request":null,"source_name":"MiniMax-M3 model card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M3"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-52be6af49920e6af5d39","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://scale.com/leaderboard/swe_bench_pro_public","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-9a626a7da4408eb8b9c49c44","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Hugging Face benchmark labels for the explicit Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Hugging Face benchmark labels for the explicit Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen/Qwen3-Coder-480B-A35B-Instruct","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":37,"rawValue":38.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cbdcc8d8029eccf3350bd0d8ce090539072d43c5","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":37,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=37;row=36","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":37,"metadata":{"author":"Qwen","filename":".eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":480154875392,"pull_request":33,"source_name":"SWE-Bench Pro official evaluation results","source_status":"reported","submission_source_url":"https://scale.com/leaderboard/swe_bench_pro_public"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":38.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-6f3c513697c2a2a2272d","canonicalModelId":"qwen/qwen3-coder-next@2026-02-03","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/papers/2603.00729","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-523dc5f8ccf5894f61ac135b","mappingCandidates":["qwen/qwen3-coder-next@2026-02-03"],"mappingEvidence":{"confidence":"high","note":"Hugging Face benchmark labels for the explicit Qwen3-Coder-Next checkpoint.","registry":"model_aliases.json"},"mappingNote":"Hugging Face benchmark labels for the explicit Qwen3-Coder-Next checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3-Coder-Next","modelRef":"Qwen/Qwen3-Coder-Next","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/.eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":34,"rawValue":44.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1fd6a85c7ce6de96622e045b9adaa1161b3782ce","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":34,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=34;row=33","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":34,"metadata":{"author":"Qwen","filename":".eval_results/.eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":79674391296,"pull_request":43,"source_name":"Qwen3-Coder-Next technical report","source_status":"reported","submission_source_url":"https://huggingface.co/papers/2603.00729"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":44.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-29539334601ea4626d2b","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-011e3d0118c678353230c777","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":24,"rawValue":53.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8ea1a40c38c51f6120a8c4d4fa302f4fdd16fb79","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":24,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=24;row=23","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":24,"metadata":{"author":"Qwen","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":53.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-2252cd3384c261f97e6a","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-7b005586b943e6b046d94393","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":28,"rawValue":49.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0817a7514c7f43694679f5ded480ee44341c4050","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":28,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=28;row=27","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":28,"metadata":{"author":"Qwen","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-e62fa29e6a4012ff73dc","canonicalModelId":"qwen/qwen3.8-2.4t-a95b@2026-08-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-fe4a28fdd63b09fcbb3d29f9","mappingCandidates":["qwen/qwen3.8-2.4t-a95b@2026-08-12"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 2.4T A95B","modelRef":"Qwen/Qwen3.8-2.4T-A95B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":1,"rawValue":67.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a88c875df01deb60f060e6e388810d5bd2ac523a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":1,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=1;row=0","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":1,"metadata":{"author":"Qwen","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":2446182725504,"pull_request":6,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-acfd6436c5addd82ddfe","canonicalModelId":"qwen/qwen3.8-27b@2026-08-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-1d3de0ab7d280e34e97ee4e5","mappingCandidates":["qwen/qwen3.8-27b@2026-08-14"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 27B","modelRef":"Qwen/Qwen3.8-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/Qwen3.8-27B.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":6,"rawValue":61.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"14f07ac549836f718f394e2933a1a3f234809ab2","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":6,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=6;row=5","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":6,"metadata":{"author":"Qwen","filename":".eval_results/Qwen3.8-27B.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":22,"source_name":"Qwen3.8-27B model card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":61.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-fdfe82dce303ee394b57","canonicalModelId":"qwen/qwen3.8-flash-next@2026-08-26","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-3d51674bf01108a25310aa3b","mappingCandidates":["qwen/qwen3.8-flash-next@2026-08-26"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.8 Flash-Next","modelRef":"Qwen/Qwen3.8-Flash-Next","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/Qwen3.8-Flash-Next.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":3,"rawValue":62.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a203b59e8ae19b706173c33c2068f57ea792169f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":3,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=3;row=2","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":3,"metadata":{"author":"Qwen","filename":".eval_results/Qwen3.8-Flash-Next.yaml","lower_is_better":false,"num_parameters":179999981459,"pull_request":12,"source_name":"Qwen3.8-Flash-Next model card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-ac1fc6fc9f4afdaa193c","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f56a19a2a86da9b25e49e0ff","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5","modelRef":"XiaomiMiMo/MiMo-V2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":19,"rawValue":56.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1771c9341aee92cfb63ef9d76bfe3c2e8d575495","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":19,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=19;row=18","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":19,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":310775040000,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-dd48293aebcc6026b29c","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d0cdb142b1beda4802d22526","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5-Pro","modelRef":"XiaomiMiMo/MiMo-V2.5-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":15,"rawValue":57.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"57a7f5a5685a79a12ae3635aee71f63bb588069b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":15,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=15;row=14","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":15,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":1023244718976,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":57.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-e27d942b1798dcdb68b0","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-de19cb17291d489c274b0765","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-ai/DeepSeek-V4-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":22,"rawValue":55.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7fce85e08aab71179ac6a503735248521274c658","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":22,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=22;row=21","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":22,"metadata":{"author":"deepseek-ai","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":1598839674782,"pull_request":110,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":55.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-ef7ed4b9f41ad4ada4c1","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-8bdbb0fe775058a4d3683096","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":26,"rawValue":50.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"54f70e4b7fd2003ce2bfa911cde7e4d32b3f7c81","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":26,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=26;row=25","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":26,"metadata":{"author":"moonshotai","filename":".eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":101,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":50.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-66e06191838d9ede465f","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2508cbed93aa0ecdca7e9b6d","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":11,"rawValue":58.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"145b3c7c1d83315ae3944f41a13d1ee84ff1dc36","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":11,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=11;row=10","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":11,"metadata":{"author":"moonshotai","filename":".eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-f230d394aecccacbd246","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-94a68adea834cbedfe447a6c","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe_bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":58.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"20b843bb0b221d5b2d0800f75e33d1e9386abb5b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":13,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":13,"metadata":{"author":"zai-org","filename":".eval_results/swe_bench_pro.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":58.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-pro","benchmarkName":"SWE-bench Pro","benchmarkVersion":null,"benchmarkVersionHint":"swebench-pro@v1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-pro-a82db8098d6555e5a529","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.2","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-69fd23cf3d3dcd826e75dc54","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.2","modelRef":"zai-org/GLM-5.2","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","protocol":{"filename":".eval_results/swe-bench_pro.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":5,"rawValue":62.1,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"82bfa6bafaae70c0652cee3ba129ebf963bcaef9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-pro/candidates.jsonl","line":5,"payloadSha256":"849df1ab0c91cce689080bcd3bc3612b64f8fb23e5290b7db80be509dacfc931","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-pro","sourceLabel":"Hugging Face · SWE-bench Pro leaderboard API","sourceLocator":"rank=5;row=4","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","sourceRow":{"artifact":"hf-swebench-pro/candidates.jsonl","line":5,"metadata":{"author":"zai-org","filename":".eval_results/swe-bench_pro.yaml","lower_is_better":false,"num_parameters":753329940480,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.2"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/ScaleAI/SWE-bench_Pro/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":62.1,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-test","benchmarkName":null,"benchmarkVersion":null,"benchmarkVersionHint":null,"benchmarkVersionId":null,"benchmarkVersionStatus":"unknown","candidateId":"cand-swebench-official-f246674c57dadce65aed","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"Sonar Foundation Agent","harnessId":"Sonar Foundation Agent","httpStatus":200,"id":"pub-66240617d9da857d880df8d3","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2025-12-19","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"Sonar Foundation Agent","leaderboard_variant":"Test","scaffold":"Sonar Foundation Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","not_independently_checked"],"rank":1,"rawValue":52.62,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"95210a1666058ebcc6fcea8166c2469f9f75bc65","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":61,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[2]=Test;results[0]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":61,"metadata":{"agent":"Sonar Foundation Agent","cost":null,"display_name":"Sonar Foundation Agent + Claude 4.5 Opus","folder":"20251219_sonar-foundation-agent_claude-opus-4-5","instance_calls":null,"logs":"s3://swe-bench-submissions/test/20251219_sonar-foundation-agent_claude-opus-4-5/logs","model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":null,"os_model":false,"os_system":false,"trajs":"s3://swe-bench-submissions/test/20251219_sonar-foundation-agent_claude-opus-4-5/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=sonar foundation agent"],"subjectType":"system","unit":"percent","value":52.62,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-12ff56020947a362b424","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"Tools","harnessId":"Tools","httpStatus":200,"id":"pub-2258bd076fb487edb11d9dfe","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"Tools","leaderboard_variant":"Verified","scaffold":"Tools","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":23,"rawValue":73.2,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e1f0047908f9ef80c9f076bb387931b1f6554c3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":107,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[22]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":107,"metadata":{"agent":"Tools","cost":null,"display_name":"Tools + Claude 4 Opus (2025-05-22)","folder":"20250522_tools_claude-4-opus","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20250522_tools_claude-4-opus/logs","model_display":"Claude 4 Opus","model_org":"Anthropic","model_release_date":null,"os_model":false,"os_system":false,"trajs":"s3://swe-bench-submissions/verified/20250522_tools_claude-4-opus/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":73.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-e252e8ab860274eb2287","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-15b826a950a00a98bfe58813","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4","modelRef":"Claude 4 Opus","observedAt":"2025-08-02","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":52,"rawValue":67.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"2e1f0047908f9ef80c9f076bb387931b1f6554c3","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":136,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[51]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":136,"metadata":{"agent":"mini-SWE-agent","cost":565.6352234999999,"display_name":"mini-SWE-agent + Claude 4 Opus (20250514)","folder":"20250802_mini-v1.0.0_claude-4-opus-20250514","instance_calls":31.076,"logs":"s3://swe-bench-submissions/verified/20250802_mini-v1.0.0_claude-4-opus-20250514/logs","model_display":"Claude 4 Opus","model_org":"Anthropic","model_release_date":20250522,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250802_mini-v1.0.0_claude-4-opus-20250514/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":67.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-8258e856e41b9ec0a0f0","canonicalModelId":"anthropic/claude-haiku-4.5@2025-10-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-2ea9de70b743b535c0706af2","mappingCandidates":["anthropic/claude-haiku-4.5@2025-10-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Haiku 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Haiku 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Haiku 4.5","modelRef":"Claude 4.5 Haiku","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":54,"rawValue":66.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9fb0f93ce1e67e60c603692a9fa9b0686b4e9e92","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":138,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[53]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":138,"metadata":{"agent":"mini-SWE-agent","cost":165.46221805000002,"display_name":"mini-SWE-agent + Claude 4.5 Haiku (high)","folder":"20260217_mini-v2.0.0_claude-4-5-haiku-high","instance_calls":66.154,"logs":null,"model_display":"Claude 4.5 Haiku","model_org":"Anthropic","model_release_date":20251015,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_claude-4-5-haiku-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-c266e9e507e860d9dd3f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"live-SWE-agent","harnessId":"live-SWE-agent","httpStatus":200,"id":"pub-2f02ba214edf8933a99e29ff","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2025-12-15","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"live-SWE-agent","leaderboard_variant":"Verified","scaffold":"live-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":1,"rawValue":79.2,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"029e66af4110085408230eea0cb7998b745e2d0a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":85,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[0]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":85,"metadata":{"agent":"live-SWE-agent","cost":null,"display_name":"live-SWE-agent + Claude 4.5 Opus medium (20251101)","folder":"20251215_livesweagent_claude-opus-4-5","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20251215_livesweagent_claude-opus-4-5/logs","model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":null,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251215_livesweagent_claude-opus-4-5/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=live-swe-agent"],"subjectType":"system","unit":"percent","value":79.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-22c9e9b821f4de34dc3f","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"Sonar Foundation Agent","harnessId":"Sonar Foundation Agent","httpStatus":200,"id":"pub-f03a7be30416643d8f320d26","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2025-12-05","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"Sonar Foundation Agent","leaderboard_variant":"Verified","scaffold":"Sonar Foundation Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":2,"rawValue":79.2,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"029e66af4110085408230eea0cb7998b745e2d0a","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":86,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[1]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":86,"metadata":{"agent":"Sonar Foundation Agent","cost":null,"display_name":"Sonar Foundation Agent + Claude 4.5 Opus","folder":"20251205_sonar-foundation-agent_claude-opus-4-5","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20251205_sonar-foundation-agent_claude-opus-4-5/logs","model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":null,"os_model":false,"os_system":false,"trajs":"s3://swe-bench-submissions/verified/20251205_sonar-foundation-agent_claude-opus-4-5/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=sonar foundation agent"],"subjectType":"system","unit":"percent","value":79.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-92f27fc4555784471558","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-f392f94d8f5e984fedf8babb","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":7,"rawValue":76.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"029e66af4110085408230eea0cb7998b745e2d0a","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":91,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[6]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":91,"metadata":{"agent":"mini-SWE-agent","cost":376.9539985,"display_name":"mini-SWE-agent + Claude 4.5 Opus (high)","folder":"20260217_mini-v2.0.0_claude-4-5-opus-high","instance_calls":32.896,"logs":null,"model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":20251124,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_claude-4-5-opus-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":76.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-6af839d27985f21163e5","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-ef34c0daeba0ed5a963d660a","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.5","modelRef":"Claude 4.5 Opus","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":20,"rawValue":74.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"029e66af4110085408230eea0cb7998b745e2d0a","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":104,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[19]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":104,"metadata":{"agent":"mini-SWE-agent","cost":360.6224985,"display_name":"mini-SWE-agent + Claude 4.5 Opus (20251101) (medium)","folder":"20251124_mini-v1.16.0_claude-opus-4-5-20251101","instance_calls":38.042,"logs":"s3://swe-bench-submissions/verified/20251124_mini-v1.16.0_claude-opus-4-5-20251101/logs","model_display":"Claude 4.5 Opus","model_org":"Anthropic","model_release_date":20251124,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251124_mini-v1.16.0_claude-opus-4-5-20251101/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":74.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-5863bd26e17f862c52ba","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"Sonar Foundation Agent","harnessId":"Sonar Foundation Agent","httpStatus":200,"id":"pub-01c961013427f3f89b07dc5d","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Sonnet 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Sonnet 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet","observedAt":"2025-11-03","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"Sonar Foundation Agent","leaderboard_variant":"Verified","scaffold":"Sonar Foundation Agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":15,"rawValue":74.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4dc585e52127c921768cded37c8f060b2d13a7f4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":99,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[14]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":99,"metadata":{"agent":"Sonar Foundation Agent","cost":null,"display_name":"Sonar Foundation Agent + Claude 4.5 Sonnet","folder":"20251103_sonar-foundation-agent_claude-sonnet-4-5","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20251103_sonar-foundation-agent_claude-sonnet-4-5/logs","model_display":"Claude 4.5 Sonnet","model_org":"Anthropic","model_release_date":null,"os_model":false,"os_system":false,"trajs":"s3://swe-bench-submissions/verified/20251103_sonar-foundation-agent_claude-sonnet-4-5/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=sonar foundation agent"],"subjectType":"system","unit":"percent","value":74.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-cb596b8ac3834bb90030","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-2a1d1d637a87a9db499bec4e","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Sonnet 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Sonnet 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":31,"rawValue":71.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4dc585e52127c921768cded37c8f060b2d13a7f4","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":115,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[30]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":115,"metadata":{"agent":"mini-SWE-agent","cost":328.9482492,"display_name":"mini-SWE-agent + Claude 4.5 Sonnet (high)","folder":"20260217_mini-v2.0.0_claude-4-5-sonnet-high","instance_calls":48.3,"logs":null,"model_display":"Claude 4.5 Sonnet","model_org":"Anthropic","model_release_date":20250929,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_claude-4-5-sonnet-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":71.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-2b6364951fec3bb2f940","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-469352f214bd034b852326ec","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Sonnet 4.5.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Sonnet 4.5.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Sonnet 4.5","modelRef":"Claude 4.5 Sonnet","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint"],"rank":40,"rawValue":70.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4dc585e52127c921768cded37c8f060b2d13a7f4","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":124,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[39]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":124,"metadata":{"agent":"mini-SWE-agent","cost":279.16737045,"display_name":"mini-SWE-agent + Claude 4.5 Sonnet (20250929)","folder":"20250929_mini-v1.13.3_sonnet-4-5-20250929","instance_calls":50.988,"logs":"s3://swe-bench-submissions/verified/20250929_mini-v1.13.3_sonnet-4-5-20250929/logs","model_display":"Claude 4.5 Sonnet","model_org":"Anthropic","model_release_date":20250929,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250929_mini-v1.13.3_sonnet-4-5-20250929/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-e08132fba10b370624ab","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-55098ff346ec5bbffa3bd24f","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of Claude Opus 4.6.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of Claude Opus 4.6.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Claude Opus 4.6","modelRef":"Claude 4.6 Opus","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":12,"rawValue":75.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d73e05d4b364f7c83080e0d21f8fd72baad86320","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":96,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[11]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":96,"metadata":{"agent":"mini-SWE-agent","cost":275.76132225,"display_name":"mini-SWE-agent + Claude 4.6 Opus","folder":"20260217_mini-v2.0.0_claude-4-6-opus","instance_calls":28.932,"logs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_claude-4-6-opus/logs","model_display":"Claude 4.6 Opus","model_org":"Anthropic","model_release_date":20260205,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_claude-4-6-opus/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":75.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-cda854b5a30b2e5cabd6","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-343c5dbd13c27471c0ca5def","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"DeepSeek V3.2","modelRef":"DeepSeek V3.2","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","not_independently_checked"],"rank":46,"rawValue":70.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49403fbeb5636ba5b024d5ee0993624eec97856d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":130,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[45]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":130,"metadata":{"agent":"mini-SWE-agent","cost":223.92324825,"display_name":"mini-SWE-agent + DeepSeek V3.2 (high)","folder":"20260217_mini-v2.0.0_deepseek-3-2-high","instance_calls":88.506,"logs":null,"model_display":"DeepSeek V3.2","model_org":"DeepSeek","model_release_date":20251201,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_deepseek-3-2-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-da094005426383959099","canonicalModelId":"deepseek/v3.2@2025-12-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3203719b90069e2e81ca76d2","mappingCandidates":["deepseek/v3.2@2025-12-01"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentation of the DeepSeek V3.2 reasoning endpoint; the harness stays on the source row.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentation of the DeepSeek V3.2 reasoning endpoint; the harness stays on the source row.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"DeepSeek V3.2","modelRef":"DeepSeek V3.2 Reasoner","observedAt":"2025-12-01","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":78,"rawValue":60.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"49403fbeb5636ba5b024d5ee0993624eec97856d","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":162,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[77]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":162,"metadata":{"agent":"mini-SWE-agent","cost":14.0381514,"display_name":"mini-SWE-agent + DeepSeek V3.2 Reasoner","folder":"20251201_mini-v1.17.1_deepseek-v3.2-reasoner","instance_calls":46.404,"logs":"s3://swe-bench-submissions/verified/20251201_mini-v1.17.1_deepseek-v3.2-reasoner/logs","model_display":"DeepSeek V3.2 Reasoner","model_org":"deepseek","model_release_date":20251201,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251201_mini-v1.17.1_deepseek-v3.2-reasoner/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":60.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-274b4d756c67e1926f7c","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-faafdc825432fe00b9e1b615","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GLM-5","modelRef":"GLM 5","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","not_independently_checked"],"rank":26,"rawValue":72.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"40d28830abcc9faba07a857a42c7957a5c538d57","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":110,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[25]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":110,"metadata":{"agent":"mini-SWE-agent","cost":267.19456811,"display_name":"mini-SWE-agent + GLM 5 (high)","folder":"20260217_mini-v2.0.0_glm-5-high","instance_calls":76.176,"logs":null,"model_display":"GLM 5","model_org":"Z-AI","model_release_date":20260211,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_glm-5-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-689f2ae1e663bbd2ad3b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-a0398be393f7a1f1210de35f","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5","modelRef":"GPT 5","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":62,"rawValue":65.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fb66f2519e79167cfe7faa5ff224f587bb11bab","selectionRank":4,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":146,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[61]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":146,"metadata":{"agent":"mini-SWE-agent","cost":140.19150875,"display_name":"mini-SWE-agent + GPT 5 (2025-08-07) (medium)","folder":"20250807_mini-v1.7.0_gpt-5","instance_calls":13.208,"logs":"s3://swe-bench-submissions/verified/20250807_mini-v1.7.0_gpt-5/logs","model_display":"GPT 5","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250807_mini-v1.7.0_gpt-5/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":65.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-c9976b9ef73928bc08f1","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-da035b58d3d556301d936aa4","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 mini","modelRef":"GPT 5 mini","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":79,"rawValue":59.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e69c314b75f7f80017118c8d2aa4af9eb6ce062","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":163,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[78]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":163,"metadata":{"agent":"mini-SWE-agent","cost":17.73853365,"display_name":"mini-SWE-agent + GPT 5 mini (2025-08-07) (medium)","folder":"20250807_mini-v1.7.0_gpt-5-mini","instance_calls":14.466,"logs":"s3://swe-bench-submissions/verified/20250807_mini-v1.7.0_gpt-5-mini/logs","model_display":"GPT 5 mini","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250807_mini-v1.7.0_gpt-5-mini/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":59.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-ad0aac31ba65b71c0a84","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-3df2cfdc1b0908e10919bbc9","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 mini","modelRef":"GPT 5 mini","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping","inferred_benchmark_version_hint","not_independently_checked"],"rank":89,"rawValue":56.2,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8e69c314b75f7f80017118c8d2aa4af9eb6ce062","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":173,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[88]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":173,"metadata":{"agent":"mini-SWE-agent","cost":23.600609549999998,"display_name":"mini-SWE-agent + GPT 5 mini","folder":"20260217_mini-v2.0.0_gpt-5-mini","instance_calls":20.342,"logs":null,"model_display":"GPT 5 mini","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_gpt-5-mini/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":56.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-04b4703818df65b9e537","canonicalModelId":"openai/gpt-5-nano@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-5b4f12f60b7d7b1afb2aee2a","mappingCandidates":["openai/gpt-5-nano@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5 nano","modelRef":"GPT 5 nano","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":147,"rawValue":34.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"390e4a6a54a3a7997407a1c92f50d63759bcee15","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":231,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[146]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":231,"metadata":{"agent":"mini-SWE-agent","cost":19.037590390000002,"display_name":"mini-SWE-agent + GPT 5 nano (2025-08-07) (medium)","folder":"20250807_mini-v1.7.0_gpt-5-nano","instance_calls":39.76,"logs":"s3://swe-bench-submissions/verified/20250807_mini-v1.7.0_gpt-5-nano/logs","model_display":"GPT 5 nano","model_org":"OpenAI","model_release_date":20250807,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250807_mini-v1.7.0_gpt-5-nano/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":34.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-4525cb014eb12a48cbaa","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-a2b95f664ba0dfccf23a0381","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","registry":"model_aliases.json"},"mappingNote":"Arena/SWE-bench presentation of GPT-5.1; setting remains in protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.1","modelRef":"GPT 5.1","observedAt":"2025-11-20","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":57,"rawValue":66.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d416b3a4bf46d834b6353b0d3e4387ce2aa626fe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":141,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[56]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":141,"metadata":{"agent":"mini-SWE-agent","cost":153.07773625000002,"display_name":"mini-SWE-agent + GPT 5.1 (2025-11-13) (medium)","folder":"20251120_mini-v1.15.0_gpt-5.1-2025-11-13","instance_calls":21.0,"logs":"s3://swe-bench-submissions/verified/20251120_mini-v1.15.0_gpt-5.1-2025-11-13/logs","model_display":"GPT 5.1","model_org":"OpenAI","model_release_date":20251112,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251120_mini-v1.15.0_gpt-5.1-2025-11-13/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-d8aad772e591301a6246","canonicalModelId":"openai/gpt-5.1-codex@2025-11-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-727fc5419de71dd9170cfdbc","mappingCandidates":["openai/gpt-5.1-codex@2025-11-12"],"mappingEvidence":{"confidence":"high","note":"Codex endpoint/harness presentation of GPT-5.1 Codex.","registry":"model_aliases.json"},"mappingNote":"Codex endpoint/harness presentation of GPT-5.1 Codex.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.1 Codex","modelRef":"GPT 5.1 Codex","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":56,"rawValue":66.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7ecdb6f675dd7263c2509472f3b1c6931ef7ae94","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":140,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[55]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":140,"metadata":{"agent":"mini-SWE-agent","cost":294.44117500000004,"display_name":"mini-SWE-agent + GPT 5.1 Codex (medium)","folder":"20251124_mini-v1.16.0_gpt-5.1-codex","instance_calls":23.882,"logs":"s3://swe-bench-submissions/verified/20251124_mini-v1.16.0_gpt-5.1-codex/logs","model_display":"GPT 5.1 Codex","model_org":"OpenAI","model_release_date":20251112,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251124_mini-v1.16.0_gpt-5.1-codex/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":66.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-18572be0dff5e871a07e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-47fb256a8486d8914e9735a2","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":27,"rawValue":72.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e9b76c66f8ae5d0670dfaf83b32f4bee68718aa","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":111,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[26]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":111,"metadata":{"agent":"mini-SWE-agent","cost":236.78350885,"display_name":"mini-SWE-agent + GPT 5.2 (high)","folder":"20260217_mini-v2.0.0_gpt-5-2-high","instance_calls":35.046,"logs":null,"model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_gpt-5-2-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-1eb7e445985265c1b02e","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-bd1260ba79ff4e833f60c7c6","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":30,"rawValue":71.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e9b76c66f8ae5d0670dfaf83b32f4bee68718aa","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":114,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[29]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":114,"metadata":{"agent":"mini-SWE-agent","cost":260.1295018,"display_name":"mini-SWE-agent + GPT 5.2 (2025-12-11) (high)","folder":"20251211_mini-v1.17.2_gpt-5.2-2025-12-11-high","instance_calls":19.762,"logs":"s3://swe-bench-submissions/verified/20251211_mini-v1.17.2_gpt-5.2-2025-12-11-high/logs","model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251211_mini-v1.17.2_gpt-5.2-2025-12-11-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":71.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":3,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-015c2810b6aec699cd99","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-ab2846599cba41047052b234","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2","modelRef":"GPT 5.2","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":49,"rawValue":69.0,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9e9b76c66f8ae5d0670dfaf83b32f4bee68718aa","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":133,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[48]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":133,"metadata":{"agent":"mini-SWE-agent","cost":134.8337487,"display_name":"mini-SWE-agent + GPT 5.2 (2025-12-11)","folder":"20251211_mini-v1.17.2_gpt-5.2-2025-12-11","instance_calls":16.27,"logs":"s3://swe-bench-submissions/verified/20251211_mini-v1.17.2_gpt-5.2-2025-12-11/logs","model_display":"GPT 5.2","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251211_mini-v1.17.2_gpt-5.2-2025-12-11/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":69.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-8aa02b70693071eb626f","canonicalModelId":"openai/gpt-5.2-codex@2025-12-18","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-fe0156c1df5b045d06f5ab75","mappingCandidates":["openai/gpt-5.2-codex@2025-12-18"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.2 Codex endpoint.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.2 Codex endpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5.2 Codex","modelRef":"GPT 5.2 Codex","observedAt":"2026-02-19","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":25,"rawValue":72.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"604bb0132339262f49bdd5b42cd14c5ef725066a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":109,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[24]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":109,"metadata":{"agent":"mini-SWE-agent","cost":224.71194935,"display_name":"mini-SWE-agent + GPT 5.2 Codex","folder":"20260219_mini-v2.0.0_gpt-5-2-codex","instance_calls":28.072,"logs":"s3://swe-bench-submissions/verified/20260219_mini-v2.0.0_gpt-5-2-codex/logs","model_display":"GPT 5.2 Codex","model_org":"OpenAI","model_release_date":20251211,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260219_mini-v2.0.0_gpt-5-2-codex/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-f22300f224f2f06ab46e","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"Prometheus-v1.2.1","harnessId":"Prometheus-v1.2.1","httpStatus":200,"id":"pub-d929351608680f75f9b2d521","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5","modelRef":"GPT-5","observedAt":"2025-10-15","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"Prometheus-v1.2.1","leaderboard_variant":"Verified","scaffold":"Prometheus-v1.2.1","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","inferred_benchmark_version_hint","not_independently_checked"],"rank":19,"rawValue":74.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fb66f2519e79167cfe7faa5ff224f587bb11bab","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":103,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[18]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":103,"metadata":{"agent":"Prometheus-v1.2.1","cost":null,"display_name":"Prometheus-v1.2.1 + GPT-5","folder":"20251015_Prometheus_v1.2.1_gpt5","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20251015_Prometheus_v1.2.1_gpt5/logs","model_display":"GPT-5","model_org":"OpenAI","model_release_date":null,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20251015_Prometheus_v1.2.1_gpt5/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":74.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-a22a43302c1430915ca4","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-94f81ac2d6c5cf97bbce25bc","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5","modelRef":"GPT-5","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"OpenHands","leaderboard_variant":"Verified","scaffold":"OpenHands","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","inferred_benchmark_version_hint"],"rank":29,"rawValue":71.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fb66f2519e79167cfe7faa5ff224f587bb11bab","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":113,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[28]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":113,"metadata":{"agent":"OpenHands","cost":null,"display_name":"OpenHands + GPT-5","folder":"20250807_openhands_gpt5","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20250807_openhands_gpt5/logs","model_display":"GPT-5","model_org":"OpenAI","model_release_date":null,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250807_openhands_gpt5/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":71.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":4,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-0f5d56c588b5ed60e0ba","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"Prometheus-v1.2","harnessId":"Prometheus-v1.2","httpStatus":200,"id":"pub-80ab4d855d41f4602922a535","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"GPT-5","modelRef":"GPT-5","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"Prometheus-v1.2","leaderboard_variant":"Verified","scaffold":"Prometheus-v1.2","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","inferred_benchmark_version_hint","not_independently_checked"],"rank":32,"rawValue":71.2,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0fb66f2519e79167cfe7faa5ff224f587bb11bab","selectionRank":3,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":116,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[31]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":116,"metadata":{"agent":"Prometheus-v1.2","cost":null,"display_name":"Prometheus-v1.2 + GPT-5","folder":"20250929_Prometheus_v1.2_gpt5","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20250929_Prometheus_v1.2_gpt5/logs","model_display":"GPT-5","model_org":"OpenAI","model_release_date":null,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250929_Prometheus_v1.2_gpt5/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system"],"subjectType":"system","unit":"percent","value":71.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-8cb44243b9e1db200c31","canonicalModelId":"google/gemini-2.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-258bb95005f11c2af16076d4","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Gemini 2.5 Flash","modelRef":"Gemini 2.5 Flash","observedAt":"2025-07-26","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","inferred_benchmark_version_hint"],"rank":155,"rawValue":28.73,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d78c4226e19fc0ed2e7e372d3906ae0479484416","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":239,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[154]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":239,"metadata":{"agent":"mini-SWE-agent","cost":66.942859175,"display_name":"mini-SWE-agent + Gemini 2.5 Flash (2025-04-17)","folder":"20250726_mini-v1.0.0_gemini-2.5-flash","instance_calls":26.646,"logs":"s3://swe-bench-submissions/verified/20250726_mini-v1.0.0_gemini-2.5-flash/logs","model_display":"Gemini 2.5 Flash","model_org":"Google DeepMind","model_release_date":20250417,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250726_mini-v1.0.0_gemini-2.5-flash/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":28.73,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-7bb8affaa001064a9234","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-b81068f5395da5f7062fed67","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Gemini 2.5 Pro","modelRef":"Gemini 2.5 Pro","observedAt":"2025-07-26","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"system"},"publishedAt":null,"qualityFlags":["agent_system_score","inferred_benchmark_version_hint"],"rank":98,"rawValue":53.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3ee98aee95642f8e4b11feab5ba6d51cdf075246","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":182,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[97]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":182,"metadata":{"agent":"mini-SWE-agent","cost":144.18604781250002,"display_name":"mini-SWE-agent + Gemini 2.5 Pro (2025-05-06)","folder":"20250726_mini-v1.0.0_gemini-2.5-pro","instance_calls":20.478,"logs":"s3://swe-bench-submissions/verified/20250726_mini-v1.0.0_gemini-2.5-pro/logs","model_display":"Gemini 2.5 Pro","model_org":"Google DeepMind","model_release_date":20250506,"os_model":false,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250726_mini-v1.0.0_gemini-2.5-pro/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"system","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=system","harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":53.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-86162cd04a2162b52e0b","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-bf73363fa04ba61329604ef4","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Kimi K2.5","modelRef":"Kimi K2.5","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":38,"rawValue":70.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"dca9cbaf844aff0e3fc0353816191559b01a0448","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":122,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[37]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":122,"metadata":{"agent":"mini-SWE-agent","cost":73.276870372,"display_name":"mini-SWE-agent + Kimi K2.5 (high)","folder":"20260217_mini-v2.0.0_kimi-k2-5-high","instance_calls":51.178,"logs":null,"model_display":"Kimi K2.5","model_org":"Moonshot AI","model_release_date":20260127,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_kimi-k2-5-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":70.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-113d5bea19239f75b84b","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-826c28218f2ba606903c3a23","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"MiniMax M2.5","modelRef":"MiniMax M2.5","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":null,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["inferred_benchmark_version_hint","not_independently_checked"],"rank":10,"rawValue":75.8,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7d36309d7bfd91751c410ced1409f04c53f40c1d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":94,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[9]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":94,"metadata":{"agent":"mini-SWE-agent","cost":36.644993922,"display_name":"mini-SWE-agent + MiniMax M2.5 (high)","folder":"20260217_mini-v2.0.0_minimax-2-5-high","instance_calls":60.45,"logs":null,"model_display":"MiniMax M2.5","model_org":"Minimax","model_release_date":20260212,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20260217_mini-v2.0.0_minimax-2-5-high/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":75.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-e540c616089541236faa","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f631fbea42a70e26423d75cb","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.1","modelRef":"MiniMaxAI/MiniMax-M2.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":28,"rawValue":74,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"efa5fbeb2dc788af32756d422e2056c5655974d3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":28,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=28;row=27","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":28,"metadata":{"author":"MiniMaxAI","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":36,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":74.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-3e39c469b33b1fea1535","canonicalModelId":"minimax/m2.5@2026-02","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.swebench.com/","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-132ffede475084ac852445fb","mappingCandidates":["minimax/m2.5@2026-02"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.5","modelRef":"MiniMaxAI/MiniMax-M2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":21,"rawValue":75.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f2502af8f0f697e4b11bb36bf3da8b7e2b1716e8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":21,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=21;row=20","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":21,"metadata":{"author":"MiniMaxAI","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":228703644928,"pull_request":null,"source_name":"SWE-Bench official evaluation","source_status":"reported","submission_source_url":"https://www.swebench.com/"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":75.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-44e2f8d2acbd8a7f8d3a","canonicalModelId":"minimax/m3@2026-05-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M3","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-c81797af459ec24e92abd2ef","mappingCandidates":["minimax/m3@2026-05-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M3","modelRef":"MiniMaxAI/MiniMax-M3","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/minimax-m3.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":6,"rawValue":80.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9eb6d0854576c2cca7e062b78a3730560407c1ea","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":6,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=6;row=5","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":6,"metadata":{"author":"MiniMaxAI","filename":".eval_results/minimax-m3.yaml","lower_is_better":false,"num_parameters":427040140160,"pull_request":null,"source_name":"MiniMax-M3 model card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M3"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-765bb0b106ecfd2cff65","canonicalModelId":"qwen/qwen3-coder-next@2026-02-03","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/papers/2603.00729","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-43feee1dab4982c4162b7cb0","mappingCandidates":["qwen/qwen3-coder-next@2026-02-03"],"mappingEvidence":{"confidence":"high","note":"Hugging Face benchmark labels for the explicit Qwen3-Coder-Next checkpoint.","registry":"model_aliases.json"},"mappingNote":"Hugging Face benchmark labels for the explicit Qwen3-Coder-Next checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3-Coder-Next","modelRef":"Qwen/Qwen3-Coder-Next","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":40,"rawValue":70.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"8b109c17fa4af663a066db891466ffa09a962918","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":40,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=40;row=39","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":40,"metadata":{"author":"Qwen","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":79674391296,"pull_request":43,"source_name":"Qwen3-Coder-Next technical report","source_status":"reported","submission_source_url":"https://huggingface.co/papers/2603.00729"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-eab5205567d7296b4039","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-fd365c4f799b5c6338fbcc64","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":35,"rawValue":72,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"13e4b175cc2d9b165b2055062609217341fd6374","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":35,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=35;row=34","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":35,"metadata":{"author":"Qwen","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":8,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-2578ab75361cfa396844","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-9c3fcaf6076c521eeffe7c43","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":33,"rawValue":72.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1772e9904efefe7bf7e1231154422a4aa91af754","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":33,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=33;row=32","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":33,"metadata":{"author":"Qwen","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-a61357c816e5823af471","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2f4352e2ac96371ae790794d","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":19,"rawValue":76.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"562678748adc1a5226c4f75f4c3768a61c8c7bd9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":19,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=19;row=18","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":19,"metadata":{"author":"Qwen","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":9,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":76.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-8dd3387258e00bdaaa98","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-610bb1943691a735b07aea80","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe-bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":18,"rawValue":77.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3ac684f6848daef8a787070a1b9ebf573791e75b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":18,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=18;row=17","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":18,"metadata":{"author":"Qwen","filename":".eval_results/swe-bench_verified.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":2,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-27236e2cf168e2b0d868","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-e4540f62b87243401968b6d0","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe-bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":31,"rawValue":73.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4053d77c798bfbc048c5977172405a9edf23b57c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":31,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=31;row=30","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":31,"metadata":{"author":"Qwen","filename":".eval_results/swe-bench_verified.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-6c51496094750875d2be","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"mini-SWE-agent","harnessId":"mini-SWE-agent","httpStatus":200,"id":"pub-b70b4a9d3726dba75675e344","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentations of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentations of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder 480B/A35B Instruct","observedAt":"2025-08-02","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":true,"harness":"mini-SWE-agent","leaderboard_variant":"Verified","scaffold":"mini-SWE-agent","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint"],"rank":91,"rawValue":55.4,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f0ee0a483edaf73bd5513ef79410950adda86af","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":175,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[90]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":175,"metadata":{"agent":"mini-SWE-agent","cost":123.963554,"display_name":"mini-SWE-agent + Qwen3-Coder 480B/A35B Instruct","folder":"20250802_mini-v1.0.0_qwen3-coder-480b-a35b-instruct","instance_calls":64.792,"logs":"s3://swe-bench-submissions/verified/20250802_mini-v1.0.0_Qwen3-Coder-480B-A35B-Instruct/logs","model_display":"Qwen3-Coder 480B/A35B Instruct","model_org":"Qwen","model_release_date":20250722,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250802_mini-v1.0.0_Qwen3-Coder-480B-A35B-Instruct/trajs"},"sourceStatus":"verified"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["harness:id=mini-swe-agent"],"subjectType":"system","unit":"percent","value":55.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-swebench-official-68f7dffc2d0ffd6b5b7c","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","harness":"OpenHands","harnessId":"OpenHands","httpStatus":200,"id":"pub-38fc0d26e0c6605142dc284b","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"SWE-bench presentations of the Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"SWE-bench presentations of the Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"resolved","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen3-Coder-480B-A35B-Instruct","observedAt":"2025-08-05","parserVersion":"0.1.0","payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","protocol":{"checked":false,"harness":"OpenHands","leaderboard_variant":"Verified","scaffold":"OpenHands","subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","not_independently_checked"],"rank":47,"rawValue":69.6,"retrievedAt":"2026-08-27T04:57:09Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f0ee0a483edaf73bd5513ef79410950adda86af","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"swebench-official/candidates.jsonl","line":131,"payloadSha256":"fa4b61d3167dfe99e1a834e007a38372c5bac07b7627f8e2c3904fb48cd4a006","retrievedAt":"2026-08-27T04:57:09Z"}],"sourceApiUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","sourceId":"swebench-official","sourceLabel":"SWE-bench · official leaderboard JSON","sourceLocator":"leaderboards[3]=Verified;results[46]","sourceMetricId":"resolved","sourcePageUrl":"https://www.swebench.com/","sourceRow":{"artifact":"swebench-official/candidates.jsonl","line":131,"metadata":{"agent":"OpenHands","cost":null,"display_name":"OpenHands + Qwen3-Coder-480B-A35B-Instruct","folder":"20250805_openhands-Qwen3-Coder-480B-A35B-Instruct","instance_calls":null,"logs":"s3://swe-bench-submissions/verified/20250805-openhands-Qwen3-Coder-480B-A35B-Instruct/logs","model_display":"Qwen3-Coder-480B-A35B-Instruct","model_org":"Qwen","model_release_date":null,"os_model":true,"os_system":true,"trajs":"s3://swe-bench-submissions/verified/20250805-openhands-Qwen3-Coder-480B-A35B-Instruct/trajs"},"sourceStatus":"reported"},"sourceSubjectType":"model","sourceUrl":"https://raw.githubusercontent.com/swe-bench/swe-bench.github.io/master/data/leaderboards.json","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":69.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-821b381b1b2fc7864a4b","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-287d106fa990a79bb917df85","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5-Pro","modelRef":"XiaomiMiMo/MiMo-V2.5-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe-bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":12,"rawValue":78.9,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1fcb1e0d473ca92d0d58288b4a8c942ef8e06fb6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":12,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=12;row=11","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":12,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/swe-bench_verified.yaml","lower_is_better":false,"num_parameters":1023244718976,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":78.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-05-22","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-7cf4430ec7d653b6209d","canonicalModelId":"anthropic/claude-opus@4","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-441234e156b65371e01dbeeb","mappingCandidates":["anthropic/claude-opus@4"],"mappingEvidence":{"confidence":"high","note":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","registry":"model_aliases.json"},"mappingNote":"Epoch labels for the dated Claude Opus 4 release; context and setting suffixes remain visible in the source protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4","modelRef":"claude-opus-4-20250514","observedAt":"2025-05-22","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7066115702479339","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1da19c4e189bfc902dc6d294d00699d3a1a3f6a9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4842,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=31;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4842,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-05-22","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":70.66115702479338,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-11-24","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-c3a46645d43a028057c1","canonicalModelId":"anthropic/claude-opus-4.5@2025-11-01","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a0d0fe0a2ec644ab2c6f86de","mappingCandidates":["anthropic/claude-opus-4.5@2025-11-01"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.5","modelRef":"claude-opus-4-5-20251101","observedAt":"2025-11-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7665289256198347","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3c900a16e35bed36463e4ed372ee32f10c5df760","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4843,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=32;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4843,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-11-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.65289256198346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-bedabcb72c687de582d4","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e3e3a2d0a7f5304cf01f07d8","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7871900826446281","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"328b03d67df9582f22ef2c4b6d62a7baa6a84dd7","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4828,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=17;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4828,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.71900826446281,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-56f900cdd5feb2991075","canonicalModelId":"anthropic/claude-opus-4.6@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-2aa011c75babc1b951800ec8","mappingCandidates":["anthropic/claude-opus-4.6@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.6","modelRef":"claude-opus-4-6","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.756198347107438","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"328b03d67df9582f22ef2c4b6d62a7baa6a84dd7","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4839,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=28;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4839,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":75.6198347107438,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-04-16","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-161c7f6ab3dcbe671b67","canonicalModelId":"anthropic/claude-opus-4.7@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f826623cbc115e64d6a920c","mappingCandidates":["anthropic/claude-opus-4.7@current"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Opus 4.7","modelRef":"claude-opus-4-7_max","observedAt":"2026-04-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.8347107438016529","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"bce6333c7d7ad8c38ac0ff2a24e148bc76caaee3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4822,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=11;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4822,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-04-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":83.47107438016529,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-09-29","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f3b2f01a6f2f57407adc","canonicalModelId":"anthropic/claude-sonnet-4.5@2025-09-29","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-83fe1205a48e3005878cf562","mappingCandidates":["anthropic/claude-sonnet-4.5@2025-09-29"],"mappingEvidence":{"confidence":"high","note":"Same dated release; effort/context/search presentation is retained as protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Same dated release; effort/context/search presentation is retained as protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.5","modelRef":"claude-sonnet-4-5-20250929","observedAt":"2025-09-29","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7128099173553719","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"cccc8a252e36d073aa7aee0dc8ed1f8b3508e1df","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4844,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=33;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4844,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2025-09-29","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.28099173553719,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-02-17","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0339a8cb0200ca78f07a","canonicalModelId":"anthropic/claude-sonnet-4.6@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4246b88f75a7a0f774add2b3","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Claude Sonnet 4.6","modelRef":"claude-sonnet-4-6","observedAt":"2026-02-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7520661157024794","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5cf5731c003010d571f372eafea2204b762c5a81","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4826,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=15;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4826,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Anthropic","release_date":"2026-02-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":75.20661157024794,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-d2e3b4777edbe98243e9","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-aaf63773eb743c9029b1b42f","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-ai/DeepSeek-V4-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe-bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":10,"rawValue":79,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"872a2db61b3e8ee18bd868605503d05ed886035d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":10,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=10;row=9","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":10,"metadata":{"author":"deepseek-ai","filename":".eval_results/swe-bench_verified.yaml","lower_is_better":false,"num_parameters":290944616402,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":79.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-b6f5504014734ee68893","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-5d4ce2443f61a1272cd4cadc","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-ai/DeepSeek-V4-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe-bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":5,"rawValue":80.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e667e13174d2a6c965117a69ba508dad1deec96e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":5,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=5;row=4","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":5,"metadata":{"author":"deepseek-ai","filename":".eval_results/swe-bench_verified.yaml","lower_is_better":false,"num_parameters":1598839674782,"pull_request":110,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-04-24","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-a470209eb0979c601a18","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-555aa442e010495a4ba6a856","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-v4-pro_max","observedAt":"2026-04-24","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7763975155279503","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"5f9b8ddf545db3ed76bab9e5de0bdafbea083fe8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4815,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=4;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4815,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"DeepSeek","release_date":"2026-04-24","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.63975155279503,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-06-17","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-11af5da8901bc749aaec","canonicalModelId":"google/gemini-2.5-pro@2025-06-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-84464497526161af34db3f5e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 2.5 Pro","modelRef":"gemini-2.5-pro","observedAt":"2025-06-17","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5755693581780539","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffa3604da6b36a13ee078005e3d535030ae76aca","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4833,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=22;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4833,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Google DeepMind","release_date":"2025-06-17","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.55693581780539,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-5325aa689bce95392b4f","canonicalModelId":"google/gemini-3.5-flash@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-ac21a30467f04740a69a8e89","mappingCandidates":["google/gemini-3.5-flash@current"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for Gemini 3.5 Flash.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for Gemini 3.5 Flash.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Gemini 3.5 Flash","modelRef":"gemini-3.5-flash_high","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7933884297520661","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1819825273200667224f3ed15f39211d84ce0676","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4816,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=5;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4816,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Google","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":79.33884297520662,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-02-11","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-f8b34ea866a8bacf6f56","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-1eb47b658dd7a02d9ad473bd","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5","modelRef":"glm-5","observedAt":"2026-02-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7207792207792207","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ace3f6c54235775c50d932d1dc30c151f1cf170a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4832,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=21;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4832,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-02-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":72.07792207792207,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-04-07","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-9f224aa2c9099ad5a810","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-52c08732904570300c88ca4e","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.1","modelRef":"glm-5.1","observedAt":"2026-04-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7417","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"6ef0e4f5213aa7df101e97b919e97b278dbc689f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4818,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=7;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4818,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-04-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.17,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-06-16","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-10c069beea8d13322ea6","canonicalModelId":"zai/glm-5.2@2026-06-16","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-e1d33c111b525aad3f139e4b","mappingCandidates":["zai/glm-5.2@2026-06-16"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GLM-5.2","modelRef":"glm-5.2_max","observedAt":"2026-06-16","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.787","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"978165d437c770b3a21cb44bdbcc6652545ff4b9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4813,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=2;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4813,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Z.ai (Zhipu AI)","release_date":"2026-06-16","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":78.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4970ca83d8da780bf81b","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-12a208fe8351327b872e98fa","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_high","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7355371900826446","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffe3824639a9ce35181968fb48d74828a7d67905","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4841,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=30;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4841,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":73.55371900826447,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-fb582dc49109889b89a3","canonicalModelId":"openai/gpt-5@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-5a090d6b131c30e9e7401ba3","mappingCandidates":["openai/gpt-5@2025-08-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5","modelRef":"gpt-5-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","heuristic_model_mapping","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7148760330578512","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ffe3824639a9ce35181968fb48d74828a7d67905","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4845,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=34;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4845,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":71.48760330578511,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-08-07","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-4ec05071b6772e71a080","canonicalModelId":"openai/gpt-5-mini@2025-08-07","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-b0b34132a6cceb51b74781b9","mappingCandidates":["openai/gpt-5-mini@2025-08-07"],"mappingEvidence":{"confidence":"high","note":"Epoch effort spellings for the dated GPT-5 mini release.","registry":"model_aliases.json"},"mappingNote":"Epoch effort spellings for the dated GPT-5 mini release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5 mini","modelRef":"gpt-5-mini-2025-08-07_medium","observedAt":"2025-08-07","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6466942148760331","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a405704a76427553b2cb7be474c9323e4d65c126","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4847,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=36;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4847,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-08-07","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":64.6694214876033,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-0cf15dfd2cb46d9245fa","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f1e54a95f079c1eeb6540a2","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6797520661157025","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f4b7534c6b58ebde19f9f1d5c7035c7198ade8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4829,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=18;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4829,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":67.97520661157024,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-11-13","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-3fcd5024e3bc0df14550","canonicalModelId":"openai/gpt-5.1@2025-11-13","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-0de55f21051348b6bd9046fb","mappingCandidates":["openai/gpt-5.1@2025-11-13"],"mappingEvidence":{"confidence":"high","note":"Effort/settings suffixes on the dated GPT-5.1 release.","registry":"model_aliases.json"},"mappingNote":"Effort/settings suffixes on the dated GPT-5.1 release.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.1","modelRef":"gpt-5.1-2025-11-13_high","observedAt":"2025-11-13","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.6590909090909091","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"a8f4b7534c6b58ebde19f9f1d5c7035c7198ade8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4831,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=20;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4831,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-11-13","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":65.9090909090909,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2025-12-11","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-83670f9c199f9cbdc3e4","canonicalModelId":"openai/gpt-5.2@2025-12-11","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a03ff70242bea3f56f5666fc","mappingCandidates":["openai/gpt-5.2@2025-12-11"],"mappingEvidence":{"confidence":"high","note":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","registry":"model_aliases.json"},"mappingNote":"Dated GPT-5.2 effort/endpoint presentations; protocol is retained.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.2","modelRef":"gpt-5.2-2025-12-11_high","observedAt":"2025-12-11","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.737603305785124","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"73f40d7266dce20cbcf3bdf07f8fe9accc21064f","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4835,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=24;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4835,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2025-12-11","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":73.7603305785124,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-02-05","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-2a19542d65e3dd7ce45e","canonicalModelId":"openai/gpt-5.3-codex@2026-02-05","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-4c47737a12434e6b16ec8231","mappingCandidates":["openai/gpt-5.3-codex@2026-02-05"],"mappingEvidence":{"confidence":"high","note":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","registry":"model_aliases.json"},"mappingNote":"Explicit GPT-5.3 Codex endpoint; effort/harness stays protocol.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.3 Codex","modelRef":"gpt-5.3-codex_high","observedAt":"2026-02-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7479338842975206","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"284e450066ec0a4fe661f1d6a83c774a02903ccc","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4824,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=13;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4824,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-02-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":74.79338842975206,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-03-05","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-687115a0d9f6fb4d8e7e","canonicalModelId":"openai/gpt-5.4@current","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-a1d51c870efd338cbdd496c6","mappingCandidates":["openai/gpt-5.4@current"],"mappingEvidence":{"confidence":"high","note":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Epoch records the explicit 2026-03-05 GPT-5.4 release with an effort suffix; effort remains in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"GPT-5.4","modelRef":"gpt-5.4-2026-03-05_high","observedAt":"2026-03-05","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.768595041322314","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"e8f38909ce4e58e528c5c96df33e5fe62d201599","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4823,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=12;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4823,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"OpenAI","release_date":"2026-03-05","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.85950413223141,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-01-27","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-576e1c4b2cd240877f01","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-d7e95c8d7a50fe23977621fd","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.5","modelRef":"kimi-k2.5","observedAt":"2026-01-27","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.737603305785124","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"c1728424e203e2d76df1495bf28aa15fed4766c0","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4830,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=19;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4830,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Moonshot","release_date":"2026-01-27","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":73.7603305785124,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-abcf1f3d55ddfd56fb6c","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-f973a5ae46b90c602cf5cee2","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Kimi K2.6","modelRef":"kimi-k2.6","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7665289256198347","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"69c672128765c49d0bceea86f36bc3b5c6760a4e","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4820,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=9;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4820,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Moonshot","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.65289256198346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-b3cb72d67da77349f998","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.swebench.com/","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-4b68987a9f33d5226493afe6","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":39,"rawValue":70.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4608dfbc5a75ccce252a73fcf79c4c208f051c4b","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":39,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=39;row=38","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":39,"metadata":{"author":"moonshotai","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":97,"source_name":"SWE-Bench official evaluation","source_status":"reported","submission_source_url":"https://www.swebench.com/"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":70.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-84e97af75ce997d5b8a0","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-eedc52e9754acd593d041ed7","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":7,"rawValue":80.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0e9fc9a4730f104c776c5fdb0328a1ad7b6b6d9c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":7,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=7;row=6","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":7,"metadata":{"author":"moonshotai","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":80.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-04-20","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-cbeb53d5a82968558aa1","canonicalModelId":"qwen/qwen3.6-max@preview","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-6b6615fe11d30021084af08f","mappingCandidates":[],"mappingStatus":"exact_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Max Preview","modelRef":"qwen3.6-max-preview","observedAt":"2026-04-20","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7665289256198347","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"7c8205742769c3a0a195c1b83eb4021c3592952a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4817,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=6;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4817,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-04-20","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":76.65289256198346,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-03-31","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-6bb760cf90d422145cc4","canonicalModelId":"qwen/qwen3.6-plus@2026-03-31","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-8f29d12dca1f1f98d1028e96","mappingCandidates":["qwen/qwen3.6-plus@2026-03-31"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.6 Plus","modelRef":"qwen3.6-plus","observedAt":"2026-03-31","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.5785123966942148","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"0c7c7443c38b0fae2fab6b3b7781dbbd39f46072","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4819,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=8;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4819,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-03-31","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":57.85123966942148,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":"2026-05-19","benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-src-epoch-benchmark-hub-702a7feb4d2769624738","canonicalModelId":"qwen/qwen3.7-max@2026-05-19","comparability":"conditional","evidenceLevel":"A","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://epoch.ai/data/benchmark_data.zip","harness":null,"harnessId":null,"httpStatus":200,"id":"pub-353f75c96a41e241a96c6a4b","mappingCandidates":["qwen/qwen3.7-max@2026-05-19"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"mean_score","modelName":"Qwen3.7 Max","modelRef":"qwen3.7-max","observedAt":"2026-05-19","parserVersion":"0.1.0","payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","protocol":{"harness":null,"subject_type":"model"},"publishedAt":null,"qualityFlags":["curated_model_alias","fraction_scaled_to_percent","inferred_benchmark_version_hint"],"rank":null,"rawValue":"0.7727272727272727","retrievedAt":"2026-08-27T08:25:00Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3256bb3d88818343dc5070782afe4d702c5c21d1","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4814,"payloadSha256":"30a9286f868f3414905cbfbf3606917e5c384a3448e17df957bccaa1eeb7a602","retrievedAt":"2026-08-27T08:25:00Z"}],"sourceApiUrl":"https://epoch.ai/data/benchmark_data.zip","sourceId":"src-epoch-benchmark-hub","sourceLabel":"Epoch AI · Benchmarking Hub downloadable snapshot","sourceLocator":"swe_bench_verified.csv:row=3;column=mean_score","sourceMetricId":"mean_score","sourcePageUrl":"https://epoch.ai/benchmarks/use-this-data","sourceRow":{"artifact":"src-epoch-benchmark-hub/candidates.jsonl","line":4814,"metadata":{"epoch_file":"swe_bench_verified.csv","notes":null,"organization":"Alibaba","release_date":"2026-05-19","source_link":null},"sourceStatus":"candidate"},"sourceSubjectType":"model","sourceUrl":"https://epoch.ai/data/benchmark_data.zip","status":"reported","subjectInferredBy":["source:subject_type=model"],"subjectType":"model","unit":"percent","value":77.27272727272727,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-61114d35bf67e1cbc42e","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-23bedbd09170c27a18458c05","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":30,"rawValue":73.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"607f29ff68af341ab2eca0461486e9d8bf2518ae","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":30,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=30;row=29","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":30,"metadata":{"author":"zai-org","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":358337791296,"pull_request":50,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":73.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-c0d8ac9a776dfcec8067","canonicalModelId":"zai/glm-4.7-flash@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-4.7-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f634e67ccde86baa0c9e7e57","mappingCandidates":["zai/glm-4.7-flash@current"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7 Flash","modelRef":"zai-org/GLM-4.7-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":59,"rawValue":59.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"58e5ca9f6fdde973b324399d88483a12b94b8861","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":59,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=59;row=58","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":59,"metadata":{"author":"zai-org","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":31221488576,"pull_request":69,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-4.7-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-eefc5d8c36d7e2406fdb","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5/","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f6cd0cfcd446a60dc5434d4b","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":15,"rawValue":77.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"beee3d909318da790cf76d80bc933ea6d9966b52","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":15,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=15;row=14","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":15,"metadata":{"author":"zai-org","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5/"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":77.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"swebench-verified","benchmarkName":"SWE-bench Verified","benchmarkVersion":null,"benchmarkVersionHint":"swebench-verified@verified","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-swebench-verified-98be1d51479210503a58","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.swebench.com/","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-bebdae5a7299c00757f819be","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","protocol":{"filename":".eval_results/swe_bench_verified.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":32,"rawValue":72.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"beee3d909318da790cf76d80bc933ea6d9966b52","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-swebench-verified/candidates.jsonl","line":32,"payloadSha256":"9d2db66b8b3c418efe70744b36f6311fed461ad78c951cbc96159cb7ab400ff8","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-swebench-verified","sourceLabel":"Hugging Face · SWE-bench Verified leaderboard API","sourceLocator":"rank=32;row=31","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","sourceRow":{"artifact":"hf-swebench-verified/candidates.jsonl","line":32,"metadata":{"author":"zai-org","filename":".eval_results/swe_bench_verified.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"SWE-Bench official evaluation","source_status":"reported","submission_source_url":"https://www.swebench.com/"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/SWE-bench/SWE-bench_Verified/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":72.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-af201ecbcde6655714f5","canonicalModelId":"minimax/m2.1@2025-12-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-09b19fb6a8ca94e8db14a39c","mappingCandidates":["minimax/m2.1@2025-12-23"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.1","modelRef":"MiniMaxAI/MiniMax-M2.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":34,"rawValue":29.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"351e68d226bff1df46a3c3fb58757d4f3152ed4d","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":34,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=34;row=33","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":34,"metadata":{"author":"MiniMaxAI","filename":".eval_results/terminal_bench.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":30,"source_name":"Terminal-Bench Leaderboard","source_status":"reported","submission_source_url":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":29.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-1d082ed62054caeb7d16","canonicalModelId":"minimax/m2.7@2026-03-18","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-0e2cafc0c4578e634f6fb919","mappingCandidates":["minimax/m2.7@2026-03-18"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiniMax M2.7","modelRef":"MiniMaxAI/MiniMax-M2.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":10,"rawValue":57,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"eeff942f3f45f10d2c65a6894bfe7b7159c36cc3","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":10,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=10;row=9","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":10,"metadata":{"author":"MiniMaxAI","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":228689764864,"pull_request":13,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":57.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-1c099533f8ebb876f970","canonicalModelId":"qwen/qwen3-coder@current","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-e84f87f42832768d2767bb0d","mappingCandidates":["qwen/qwen3-coder@current"],"mappingEvidence":{"confidence":"high","note":"Hugging Face benchmark labels for the explicit Qwen3-Coder 480B-A35B Instruct checkpoint.","registry":"model_aliases.json"},"mappingNote":"Hugging Face benchmark labels for the explicit Qwen3-Coder 480B-A35B Instruct checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3-Coder 480B-A35B","modelRef":"Qwen/Qwen3-Coder-480B-A35B-Instruct","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":38,"rawValue":23.9,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ccc773573aaec1f5be64d74285cbc3e9aa3ae944","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":38,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=38;row=37","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":38,"metadata":{"author":"Qwen","filename":".eval_results/terminal_bench.yaml","lower_is_better":false,"num_parameters":480154875392,"pull_request":32,"source_name":"Terminal-Bench Leaderboard","source_status":"reported","submission_source_url":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":23.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-3e61e531803600d7c214","canonicalModelId":"qwen/qwen3-coder-next@2026-02-03","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3-Coder-Next","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2d522fec927e539e48698248","mappingCandidates":["qwen/qwen3-coder-next@2026-02-03"],"mappingEvidence":{"confidence":"high","note":"Hugging Face benchmark labels for the explicit Qwen3-Coder-Next checkpoint.","registry":"model_aliases.json"},"mappingNote":"Hugging Face benchmark labels for the explicit Qwen3-Coder-Next checkpoint.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3-Coder-Next","modelRef":"Qwen/Qwen3-Coder-Next","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":22,"rawValue":36.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"166ab3f3f44f1aeec5201b59fe24b37896a9b7c9","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":22,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=22;row=21","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":22,"metadata":{"author":"Qwen","filename":".eval_results/terminal_bench_2.yaml","lower_is_better":false,"num_parameters":79674391296,"pull_request":43,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3-Coder-Next"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":36.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-41d5b23a5977e1776b1b","canonicalModelId":"qwen/qwen3.5-122b-a10b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-c36c5468c4676c1dc6a30cde","mappingCandidates":["qwen/qwen3.5-122b-a10b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 122B A10B","modelRef":"Qwen/Qwen3.5-122B-A10B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":17,"rawValue":49.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"464b0fbd7596b2d1b62d1f2284c3a97cfaed646a","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":17,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=17;row=16","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":17,"metadata":{"author":"Qwen","filename":".eval_results/terminal-bench-2_2.yaml","lower_is_better":false,"num_parameters":125086497008,"pull_request":8,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":49.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-d62ea7a2dc2fa9d55e7a","canonicalModelId":"qwen/qwen3.5-27b@2026-02-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-bc71e9455400af5bbcb90d78","mappingCandidates":["qwen/qwen3.5-27b@2026-02-24"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 27B","modelRef":"Qwen/Qwen3.5-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0_4.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":20,"rawValue":41.6,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1000a6c94996365d991794fbf837d59721117602","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":20,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=20;row=19","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":20,"metadata":{"author":"Qwen","filename":".eval_results/terminal-bench-2.0_4.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":41.6,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-884f65740ac422be87db","canonicalModelId":"qwen/qwen3.5-397b-a17b@2026-02-13","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-994efcc0153c2221f3fcddd5","mappingCandidates":["qwen/qwen3.5-397b-a17b@2026-02-13"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.5 397B A17B","modelRef":"Qwen/Qwen3.5-397B-A17B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminalbench_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":13,"rawValue":52.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"d9cce69906fe4a8df336a2a8239dd73f1678059c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":13,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=13;row=12","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":13,"metadata":{"author":"Qwen","filename":".eval_results/terminalbench_2.yaml","lower_is_better":false,"num_parameters":403397928944,"pull_request":9,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-2a928d3ac4263a9291af","canonicalModelId":"qwen/qwen3.6-27b@2026-04-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-27B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-d107962c9d82877d516d8a62","mappingCandidates":["qwen/qwen3.6-27b@2026-04-22"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 27B","modelRef":"Qwen/Qwen3.6-27B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":9,"rawValue":59.3,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"4b0916201c31b002228f7b81c8b3a95a193025e6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":9,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=9;row=8","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":9,"metadata":{"author":"Qwen","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":27781427952,"pull_request":45,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-27B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":59.3,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-60ac9b0489cd44091111","canonicalModelId":"qwen/qwen3.6-35b-a3b@2026-04-14","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-f762ecb0f96fd55823cbd164","mappingCandidates":["qwen/qwen3.6-35b-a3b@2026-04-14"],"mappingEvidence":{"confidence":"high","note":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","registry":"model_aliases.json"},"mappingNote":"Explicit Qwen release/repository name; size and variant remain in the raw source spelling.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Qwen3.6 35B A3B","modelRef":"Qwen/Qwen3.6-35B-A3B","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":15,"rawValue":51.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"89647187d4ab46ddd828ec8d5052c2fbf5a4bdf6","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":15,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=15;row=14","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":15,"metadata":{"author":"Qwen","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":35951822704,"pull_request":3,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":51.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-9c1afe04f1d20d4b7603","canonicalModelId":"xiaomi/mimo-v2.5@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-787330cefc9f3c4eb899140e","mappingCandidates":["xiaomi/mimo-v2.5@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5","modelRef":"XiaomiMiMo/MiMo-V2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":7,"rawValue":65.8,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"733aea2795262f25a8a8a27dd562829774589995","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":7,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=7;row=6","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":7,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":310775040000,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":65.8,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-1205d858047d0187c00f","canonicalModelId":"xiaomi/mimo-v2.5-pro@2026-04-23","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-21d6c77a10f527526f00c859","mappingCandidates":["xiaomi/mimo-v2.5-pro@2026-04-23"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"MiMo-V2.5-Pro","modelRef":"XiaomiMiMo/MiMo-V2.5-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":3,"rawValue":68.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ed84fdfdfa57ea2868b7cfab716aa4c31f030271","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":3,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=3;row=2","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":3,"metadata":{"author":"XiaomiMiMo","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":1023244718976,"pull_request":1,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":68.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-cb85389e98b62d786870","canonicalModelId":"deepseek/v4-flash@2026-07-31","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-a8aa0f1936041fbc4046a907","mappingCandidates":["deepseek/v4-flash@2026-07-31"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Flash","modelRef":"deepseek-ai/DeepSeek-V4-Flash","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":11,"rawValue":56.9,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"25210aa14125ee8bc7ebc94d40d7cc9833616cd4","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":11,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=11;row=10","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":11,"metadata":{"author":"deepseek-ai","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":290944616402,"pull_request":11,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":56.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-b8f646de50d880960e4c","canonicalModelId":"deepseek/v4-pro@2026-04-24","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-6b383488a5beac3b1b56ffe9","mappingCandidates":["deepseek/v4-pro@2026-04-24"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"DeepSeek V4 Pro","modelRef":"deepseek-ai/DeepSeek-V4-Pro","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal-bench-2.0.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":4,"rawValue":67.9,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"3299fba7fc3f7f89f4c4671f265031c2588f892c","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":4,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=4;row=3","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":4,"metadata":{"author":"deepseek-ai","filename":".eval_results/terminal-bench-2.0.yaml","lower_is_better":false,"num_parameters":1598839674782,"pull_request":110,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":67.9,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-bb065eefce1f1603803e","canonicalModelId":"moonshot/kimi-k2.5@2026-01-27","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-5f7c0efaf2d0252ec1ac2f87","mappingCandidates":["moonshot/kimi-k2.5@2026-01-27"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.5","modelRef":"moonshotai/Kimi-K2.5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":19,"rawValue":43.2,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"f6e108fcc4214c22a7be98a0910fb0788cbd2c78","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":19,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=19;row=18","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":19,"metadata":{"author":"moonshotai","filename":".eval_results/terminal_bench.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":90,"source_name":"Terminal-Bench Leaderboard","source_status":"reported","submission_source_url":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":43.2,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-cbe81a7e3289e41211a4","canonicalModelId":"moonshot/kimi-k2.6@2026-04-20","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/moonshotai/Kimi-K2.6","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-be2abf33b38a4c20ee1904ee","mappingCandidates":["moonshot/kimi-k2.6@2026-04-20"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"Kimi K2.6","modelRef":"moonshotai/Kimi-K2.6","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":5,"rawValue":66.7,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"9202715a5fa3ac3ad16d005f35248e4742da87da","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":5,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=5;row=4","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":5,"metadata":{"author":"moonshotai","filename":".eval_results/terminal_bench_2.yaml","lower_is_better":false,"num_parameters":1026879376368,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/moonshotai/Kimi-K2.6"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":66.7,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-3ea99cab2a3d4313636c","canonicalModelId":"zai/glm-4.7@2025-12-22","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-fcd8dc973ecd35b433ba160c","mappingCandidates":["zai/glm-4.7@2025-12-22"],"mappingEvidence":{"confidence":"high","note":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","registry":"model_aliases.json"},"mappingNote":"Explicit source model name; effort, harness and deployment mode remain in protocol metadata.","mappingStatus":"curated_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-4.7","modelRef":"zai-org/GLM-4.7","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["curated_model_alias","inferred_benchmark_version_hint","unverified_submission"],"rank":26,"rawValue":33.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"23808c49c42a4f513bae869a625760624a2cdebe","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":26,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=26;row=25","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":26,"metadata":{"author":"zai-org","filename":".eval_results/terminal_bench.yaml","lower_is_better":false,"num_parameters":358337791296,"pull_request":44,"source_name":"Terminal-Bench Leaderboard","source_status":"reported","submission_source_url":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":33.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":1,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-3444e201d050f11ebc69","canonicalModelId":"zai/glm-5@2026-02-12","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://www.tbench.ai/leaderboard/terminal-bench/2.0","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-bd4d5b802924c77f8e53698c","mappingCandidates":["zai/glm-5@2026-02-12"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5","modelRef":"zai-org/GLM-5","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":14,"rawValue":52.4,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"ab7e1c19591ef3f514c31017af242186459f7206","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":14,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=14;row=13","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":14,"metadata":{"author":"zai-org","filename":".eval_results/terminal_bench_2.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Terminal-Bench Leaderboard","source_status":"reported","submission_source_url":"https://www.tbench.ai/leaderboard/terminal-bench/2.0"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":52.4,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-5f24fe4c703e8836e78c","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-0d91ab2703545651dcdc874c","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench_2_claudecode.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":2,"rawValue":69,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1fec9b5f09296b60426aedfbc58846081c8407d8","selectionRank":1,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":2,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=2;row=1","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":2,"metadata":{"author":"zai-org","filename":".eval_results/terminal_bench_2_claudecode.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":69.0,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}
{"alternativesCount":2,"benchmarkId":"terminal-bench","benchmarkName":"Terminal-Bench 2.1","benchmarkVersion":null,"benchmarkVersionHint":"terminal-bench@2.1","benchmarkVersionId":null,"benchmarkVersionStatus":"catalog_default_hint","candidateId":"cand-hf-terminal-bench-29d4b31b9d09978dd215","canonicalModelId":"zai/glm-5.1@2026-04-07","comparability":"conditional","evidenceLevel":"C","evidenceNote":"公开来源报告值；尚未由本项目独立复现/核验。","evidenceUrl":"https://huggingface.co/zai-org/GLM-5.1","harness":"huggingface-eval-results","harnessId":"huggingface-eval-results","httpStatus":200,"id":"pub-2775fcf7296963af75b0c764","mappingCandidates":["zai/glm-5.1@2026-04-07"],"mappingStatus":"heuristic_alias","matrixExcluded":false,"matrixExcludedReason":null,"metricId":"score","modelName":"GLM-5.1","modelRef":"zai-org/GLM-5.1","observedAt":null,"parserVersion":"0.1.0","payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","protocol":{"filename":".eval_results/terminal_bench_2.yaml","harness":"huggingface-eval-results","verified":false},"publishedAt":null,"qualityFlags":["heuristic_model_mapping","inferred_benchmark_version_hint","unverified_submission"],"rank":8,"rawValue":63.5,"retrievedAt":"2026-08-27T06:23:04Z","reviewStatus":"unreviewed","selection":"curated","selectionKey":"1fec9b5f09296b60426aedfbc58846081c8407d8","selectionRank":2,"selectionReason":["canonical_model_mapping","all_source_variants"],"snapshotCount":1,"snapshotLocations":[{"artifact":"hf-terminal-bench/candidates.jsonl","line":8,"payloadSha256":"a58848d66cc6d2949246bbf78edc8327e8611019db601e3603a2238da1ee2b74","retrievedAt":"2026-08-27T06:23:04Z"}],"sourceApiUrl":null,"sourceId":"hf-terminal-bench","sourceLabel":"Hugging Face · Terminal-Bench leaderboard API","sourceLocator":"rank=8;row=7","sourceMetricId":"score","sourcePageUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","sourceRow":{"artifact":"hf-terminal-bench/candidates.jsonl","line":8,"metadata":{"author":"zai-org","filename":".eval_results/terminal_bench_2.yaml","lower_is_better":false,"num_parameters":753864139008,"pull_request":null,"source_name":"Model Card","source_status":"reported","submission_source_url":"https://huggingface.co/zai-org/GLM-5.1"},"sourceStatus":"reported"},"sourceSubjectType":null,"sourceUrl":"https://huggingface.co/api/datasets/harborframework/terminal-bench-2.0/leaderboard","status":"reported","subjectInferredBy":["default:model"],"subjectType":"model","unit":"percent","value":63.5,"verificationStatus":"not_reproduced","verification_status":"not_reproduced","verified":false}